Unnamed: 0
int64
0
378k
id
int64
49.9k
73.8M
title
stringlengths
15
150
question
stringlengths
37
64.2k
answer
stringlengths
37
44.1k
tags
stringlengths
5
106
score
int64
-10
5.87k
17,000
45,842,932
Pandas drop duplicates with partially completed data in each row and combine data
<p>I have a dataframe with duplicate IDs but the data is partially completed in multiple areas. </p> <pre><code>df = pd.DataFrame([[1234, 'Customer A', '123 Street', np.nan, np.nan], [1234, 'Customer A', np.nan, '333 Street', np.nan], [1234, 'Customer A', '12345 Street', np.nan, np.nan], ...
<p>Here's one approach using <code>apply</code> and create new columns, using <code>dict</code> creation for <code>pd.Series</code></p> <pre><code>In [1057]: cols = ['Billing Address', 'Shipping Address'] In [1058]: (df.groupby(['ID', 'Customer']) .apply(lambda g: pd.Series({'%s %s' % (x, i+1): v[x] ...
python|pandas
3
17,001
23,313,260
Do Python functions copy the input parameters to the function?
<p>I have the following example in Python:</p> <pre><code>import numpy as np import timeit # array size m = 3000 # square array a = np.random.rand(m, m) # column vector b = np.random.rand(m) # solve ti = timeit.default_timer() x = np.linalg.solve(a, b) tf = timeit.default_timer() # solve function def solver(a, b)...
<p>The speed difference is likely due to some memory caching of the contents of <code>a</code> and <code>b</code> that the second call to <code>linalg.solve</code> takes advantage of.</p> <p>And to answer the question, objects in python are passed by reference, so the only differencein the cost of passing arguments i...
python|python-3.x|numpy|timeit
1
17,002
11,707,586
How do I expand the output display to see more columns of a Pandas DataFrame?
<p>Is there a way to widen the display of output in either interactive or script-execution mode?</p> <p>Specifically, I am using the <code>describe()</code> function on a Pandas <code>DataFrame</code>. When the <code>DataFrame</code> is five columns (labels) wide, I get the descriptive statistics that I want. However...
<p><strong>Update: Pandas 0.23.4 onwards</strong></p> <p>This is not necessary. Pandas autodetects the size of your terminal window if you set <code>pd.options.display.width = 0</code>. (For older versions see at bottom.)</p> <p><code>pandas.set_printoptions(...)</code> is deprecated. Instead, use <code>pandas.set_opti...
python|pandas|printing|column-width
1,300
17,003
28,466,772
Filling date gaps in pandas dataframe
<p>I have Pandas DataFrame (loaded from .csv) with Date-time as index.. where there is/have-to-be one entry per day. The problem is that I have gaps i.e. there is days for which I have no data at all. What is the easiest way to insert rows (days) in the gaps ? Also is there a way to control what is inserted in the colu...
<p>You'll could resample by day e.g. using mean if there are multiple entries per day:</p> <pre><code>df.resample('D', how='mean') </code></pre> <p>You can then <code>ffill</code> to replace NaNs with the previous days result.</p> <p><em>See <a href="http://pandas.pydata.org/pandas-docs/stable/timeseries.html#up-and...
python|datetime|csv|pandas|pad
12
17,004
50,779,252
Creating a numpy array with dimension, mean, and variance as parameters?
<p>How do I create a numpy array with the dimensions, mean, and variance as parameters? I see there is a <code>numpy.random.randn</code> function which allows the user to specify dimensions, but that function assumes a mean of 0 and variance of 1.</p> <p>I am okay with the mean 0 part, but I want to be able to specify...
<p>You may be looking for <a href="https://docs.scipy.org/doc/numpy-1.13.0/reference/generated/numpy.random.normal.html#numpy-random-normal" rel="nofollow noreferrer"><code>numpy.random.normal</code></a>. For example:</p> <pre><code>import numpy as np arr = np.random.normal(loc=1, scale=0.50, size=(500, 500)) print(...
python|arrays|numpy|random
4
17,005
50,854,611
Pandas groupby multiple columns, sum and append answer as new column to the original dataframe
<p>I have the following dataframe:</p> <pre><code>In [1]: import pandas as pd pd.DataFrame({"AAA":["x1","x1","x1","x1"], "BBB":["y1","y1","y1","y2"], "CCC":["t1","t2","t3","t1"], "DDD":[10,11,18,17]}) Out[1]: AAA BBB CCC DDD 0 x1 y1 t1 10 1 x1 y1 t2 11 2 x1 y1...
<p>One way is to use:</p> <pre><code>df['AAA_BBB sum'] = df.groupby(['AAA', 'BBB'])['DDD'].transform(lambda x: x.sum()) </code></pre> <p>This gives:</p> <pre><code> AAA BBB CCC DDD AAA_BBB sum 0 x1 y1 t1 10 39 1 x1 y1 t2 11 39 2 x1 y1 t3 18 39 3 x1 ...
python|pandas|dataframe
5
17,006
50,784,144
How to set iteration depth when interating numpy array with e.g. nditer?
<p>supose I have the following array, representing the structure of an rgb image:</p> <pre><code>[[[ 0, 1, 2], [ 3, 4, 5]] [[ 6, 7, 8], [ 9, 10, 11]] [[12, 13, 14], [15, 16, 17]]] </code></pre> <p>How can I iterate over the pixels, e.g. [0, 1, 2] then [3, 4, 5], and receive the coresponding index? With numpys...
<p>If I understood your question correctly you could just use a simple nested loop</p> <pre><code>A = np.array([[[ 0, 1, 2], [ 3, 4, 5]], [[ 6, 7, 8], [ 9, 10, 11]], [[12, 13, 14], [15, 16, 17]]]) for i in range(A.shape[0]): for j in range(A.shape[1]): print(i, j, A[i,j,....
python|numpy|opencv
4
17,007
50,998,868
Implement MSSQL's partition by windowed clause in Pandas
<p>I’m in the process of moving a <code>MSSQL</code> database to <code>MYSQL</code> and have decided to move some stored procedures to <code>Python</code> rather than rewrite in <code>MYSQL</code>. I am using Pandas 0.23 on Python 3.5.4.</p> <p>The old <code>MSSQL</code> base uses a number of windowed functions. So fa...
<p>I am pretty sure the transform function can help.</p> <pre><code>df.groupby('Col1'')['Val'].transform(lambda x: x.rolling(3, 2).mean()) </code></pre> <p>where e.g. the value 3 is the step of the rolling window, and 2 is the minimum number of periods.</p> <p>(Just don't forget to sort your data frame before applyi...
python|sql-server|pandas
3
17,008
20,441,335
Python: "IndexError: invalid index into a 0-size array"
<p>Obviously, there are loads of threads concerning index errors. But I couldn't find one that helped me out.</p> <p>I use numpy.loadtxt to read in a function f(a,b).</p> <pre><code>a, b, f = np.loadtxt(filename, delimiter=' ', usecols=(0,1,2), unpack=True) </code></pre> <p>To get a 2d plot I found a solution in ano...
<p>Thanks to seberg's advice, I found the mistake (and I also found old stackoverflow account): </p> <pre><code>a, b, f = np.loadtxt(filename, delimiter=' ', usecols=(0,1,2), unpack=True) </code></pre> <p>I imported the wrong columns from the file. One happened to be constant. So I got nrows=0. And grid was the 0-siz...
python|arrays|numpy|indexing
1
17,009
66,657,701
Insert values in list in DataFrame cell based on string in another cell
<p>I have a pandas DataFrame df containing three columns where one is largely empty. In the first column I have a list of agreements where some of them have similar names. I want to insert a list of values in the empty column if the value in the agreements column starts with a certain string.</p> <p>Example:</p> <div c...
<p>use a regex with <code>^</code> which checks for a pattern at the start of a string.</p> <pre><code>df.loc[df['Agreement'].str.contains('^Agreement'),'Alias'] = [['Agree','Agreement']] Agreement Alias 0 SampleAgreement None 1 Agreement1 [Agree, Agreement] 2 Agreeme...
python|pandas|dataframe|startswith
0
17,010
66,506,636
create 1-D numpy.ndarray instead of 2 dimensional
<p>I am trying to do a classification using Python. I have some input columns (let k variables) and one output column.</p> <pre><code>Let inputfeatures = array([[ 0, 0, 0, ..., 0, 0, 0], [ 1, 0, 0, ..., 0, 0, 0], [ 1, 0, 0, ..., 0, 0, 0], ..., [ 0, 0, 0, ..., ...
<p>You can't simply transform an <code>n</code>-dimensional label (or &quot;target output&quot;) into a 1-dimensional one. In some cases where the output distribution is an <code>m</code>-dimensional manifold embedded into an <code>n</code>-dimensional space, you may try to do a projection first (including, if necessa...
python|arrays|numpy|classification
1
17,011
57,319,192
parsing tsv file starting with quotation mark
<p>I tried to parse a TSV file which contains two columns, some lines only contain quotation mark. Is there a way to parse them as separate lines using python without adding '\' before the quotation mark?</p> <pre><code>presents O it O in O " O classical O " O principles O on O which O ' O the O o...
<p>You can tell <code>csv.reader</code> to ignore quote characters by adding <code>quoting = csv.QUOTE_NONE</code> when you create the instance:</p> <pre><code>import csv with open("sample.tsv") as tsvfile: tsvreader = csv.reader(tsvfile, delimiter="\t", quoting=csv.QUOTE_NONE) for line in tsvreader: p...
python|pandas|csv
2
17,012
24,228,905
Sample data from combination of two probability distributions
<p>I want to make a mock catalogue. I have access to two sets of real data and I want to use their properties to generate the mock catalogue: The <strong>first</strong> one contains the information from <strong>magnitude</strong> and <strong>redshift</strong> (<code>z</code>). The <strong>second</strong> set has inf...
<p>Let R = red shift, M = magnitude, W = weight. If I understand correctly, you are trying to sample from P(W, R).</p> <p>Given the data at hand, you do not have enough information for a complete solution: note that P(W, R) = P(W | R) P(R) where P(W | R) = \int P(W, M | R) dM = \int P(W | M, R) P(M | R) dM. You can der...
numpy|statistics|scipy|probability-theory
2
17,013
43,502,319
barplot x axis construction from data pandas seaborn python
<p>So i'm trying to create a barplot using seaborn. My data is in the form</p> <pre><code>Packet number,Flavour,Contents 1,orange,4 2,orange,3 3,orange,2 4,orange,4 ... 36, orange,3 1, coffee,5 2, coffee,3 ... 1, raisin,4 etc. </code></pre> <p>My code is currently:</p> <pre><code>revels_data = pd.read_csv("testtt.tx...
<p>You want to use <code>hue</code> for that. As well, currently you are displaying the mean of each category. To calculate different function you can use <code>estimator</code>.</p> <p>Thus, your code should be:</p> <pre><code>ax = sns.barplot(x="Packet number", y="Contents", hue="Flavour", data=rd) </code></pre> <...
python|pandas|bar-chart|seaborn
2
17,014
43,552,993
How to number a columns name pandas
<p>I want to name columns in a pandas dataframe from 1 to length of the data. </p> <p>I have</p> <pre><code>foo = pd.DataFrame(n, columns=('protein', 1...n )) </code></pre>
<p>The default column names in pandas is 0 to range(len(data))-1. So if you don't pass the columns = parameter in pd.DataFrame and then rename first column, you will get desired result</p> <p>Eg: n = np.ones(10).reshape(2,5) df = pd.DataFrame(n) </p> <pre><code> 0 1 2 3 4 0 1.0 1.0 1.0 1.0 1.0 ...
python|list|pandas|dataframe
1
17,015
73,064,783
How to round values in a column of sets within a pandas dataframe?
<p>A confidence interval calculation returns two numbers inside a set of brackets, which I put into a dataframe. From this point, I need to round down the values, and put the output into an excel file.</p> <p>In the below sample code, how can I round down the values in Col1, to 2 decimals.</p> <pre><code>import pandas ...
<p>You can use <a href="https://pandas.pydata.org/docs/reference/api/pandas.DataFrame.apply.html" rel="nofollow noreferrer"><code>panda.apply</code></a> on column <code>Col1</code> and <code>round(num, 2)</code> for rounding nums to 2 decimals.</p> <pre><code>df['Col1'] = df['Col1'].apply(lambda x : tuple([round(x[0], ...
python|pandas|list|xlwings
2
17,016
72,850,621
Weights and Bias dimensions in TensorFlow for LSTM
<p>I am confused about the dimensions of the hidden weight array for an LSTM. I understand the input weight and bias dimension just fine. Now I just recently started learning about RNNs and LSTMs so maybe I do not fully understand their operation. Here is my understanding. A LSTM layer has a number of cells which hold ...
<p>Here are the relevant equations from the Wiki on <a href="https://en.wikipedia.org/wiki/Long_short-term_memory" rel="nofollow noreferrer">LSTM</a></p> <p><a href="https://i.stack.imgur.com/L6W94.png" rel="nofollow noreferrer"><img src="https://i.stack.imgur.com/L6W94.png" alt="" /></a></p> <p>Notice, that as you sai...
python|tensorflow|keras|lstm|recurrent-neural-network
2
17,017
73,082,256
How to create a list of tokenized words from dataframe column using spaCy?
<p>I'm trying to apply <code>spaCy</code>s tokenizer on dataframe column to get a new column containing list of tokens. Assume we have the following dataframe:</p> <pre><code>import pandas as pd details = { 'Text_id' : [23, 21, 22, 21], 'Text' : ['All roads lead to Rome', 'All work and no play ma...
<p>Try this</p> <pre><code>example_df[&quot;tokens&quot;] = example_df[&quot;Text&quot;].apply(lambda x : [token.text for token in nlp.tokenizer(x)]) </code></pre> <p>which gives us</p> <p><a href="https://i.stack.imgur.com/juZxN.png" rel="nofollow noreferrer"><img src="https://i.stack.imgur.com/juZxN.png" alt="enter i...
python|pandas|nlp|spacy|tokenize
1
17,018
73,072,257
Resolve warning "A NumPy version >=1.16.5 and <1.23.0 is required for this version of SciPy"?
<p>When I import SciPy or a library dependent on it, I receive the following warning message:</p> <pre><code>UserWarning: A NumPy version &gt;=1.16.5 and &lt;1.23.0 is required for this version of SciPy (detected version 1.23.1 </code></pre> <p>It's true that I am running NumPy version 1.23.1, however this message is a...
<p>I have the same issue.</p> <p>The <a href="https://docs.scipy.org/doc/scipy/dev/toolchain.html#numpy" rel="noreferrer">scipy 1.7.3 docs</a> specifies <code>1.16.5 &lt;= numpy &lt;1.24.0</code> while in scipy 1.7.3 code <a href="https://github.com/scipy/scipy/blob/59e6539cf80dc04b16b0f0ab52343381f0a7a2fa/setup.py#L55...
python|numpy|scipy|conda
5
17,019
72,924,655
Python How to replace values in specific columns (defined by an array) with zero
<p>I'm trying to replace values in specific columns with zero with python, and the column numbers are specified in another array.</p> <p>Given the following 2 numpy arrays</p> <pre><code>a = np.array([[ 1, 2, 3, 4], [ 1, 2, 1, 2], [ 0, 3, 2, 2]]) </code></pre> <p>and</p>...
<p>Your question is:</p> <blockquote> <p>I'm trying to replace values in specific columns with zero with python, and the column numbers are specified in another array.</p> </blockquote> <p>This can be done like this:</p> <pre class="lang-py prettyprint-override"><code>a[:,b] = 0 </code></pre> <p>Output:</p> <pre><code>...
python|arrays|numpy
2
17,020
70,655,623
Multiply 1 Dataframe by a row in another one selected based on its index value
<p>I am pulling my hair on this one.</p> <p>I have 2 Dataframes:</p> <p>df1 holds data for Players with their position (zone) (Forward, Midfield or Defender) and some of their</p> <p>game stats.</p> <pre><code>df1 = pd.DataFrame({'Zone': ['DEF', 'MID', 'FWD'], 'Tackles': [5, 10, 5], 'Goals': [0, 1, 1], 'Shots': [10, 5,...
<p>Append temporary <code>Zone</code> as index of <code>df1</code>:</p> <pre><code>df1['Index'] = df1.set_index('Zone', append=True).mul(df2, level=1).sum(axis=1).values print(df1) # Output Zone Tackles Goals Shots Index Player A DEF 5 0 10 30 Player B MID 10 1 5 ...
python|pandas|dataframe|indexing|.loc
0
17,021
70,632,285
Is it possible to apply some transformation ( augmentation) of Text data Dynamically during training in Tensorflow?
<p>Let us suppose I have this very simple pipeline <strong>which already works</strong>:</p> <pre><code>X = [['she let the balloon float up into the air with her hopes and dreams'], ['the old rusted farm equipment surrounded the house predicting its demise'], ['he was so preoccupied with whether or not ...
<p>When I train my models I use the map function in the input pipeline of my dataset. Here is a guide: <a href="https://www.tensorflow.org/api_docs/python/tf/data/Dataset#map" rel="nofollow noreferrer">https://www.tensorflow.org/api_docs/python/tf/data/Dataset#map</a></p> <p>You need to define a function with the trans...
tensorflow|keras|deep-learning|nlp|tensorflow2.0
0
17,022
70,698,487
How to get a Pytorch data loader per class?
<p>I want to train my model on 1 MNIST class at a time.</p> <p>I can load the data with a general loader:</p> <pre><code>import torch import torch.nn as nn import torch.optim as optim import torchvision.datasets as datasets import torchvision.transforms as transforms from torch.autograd import Variable trans = transf...
<p>A rather simple solution would involve grouping the dataset by truth value, and creating a unique dataloader per group:</p> <pre class="lang-py prettyprint-override"><code>... from torch.utils.data import Subset, DataLoader subsets = {target: Subset(train_set, [i for i, (x, y) in enumerate(train_set) if y == target...
python|pytorch|pytorch-dataloader
1
17,023
42,883,415
How to replace a value within a tensor by indices?
<p>The below code add something to a specific location within a tensor by indices (thanks to @mrry's answer <a href="https://stackoverflow.com/questions/34685947/adjust-single-value-within-tensor-tensorflow">here</a>). </p> <pre><code>indices = [[1, 1]] # A list of coordinates to update. values = [1.0] # A list of v...
<p>A simple option to updating a Tensor based on its own values or indices is using <code>tf.where</code> and <code>tf.tensor_scatter_nd_update</code>:</p> <pre class="lang-py prettyprint-override"><code>import tensorflow as tf x = tf.constant([[4.0, 43.0, 45.0], [2.0, 22.0, 6664.0], ...
python|tensorflow|tensor
3
17,024
42,770,605
choosing certain columns of a dataframe from a file
<p>I have a dataframe df</p> <pre><code>Fruit Apple Orange Banana Pear basket1 0 1 10 15 basket2 1 5 7 10 basket3 10 15 0 0 </code></pre> <p>I have another dataframe select</p> <pre><code>type1 type2 type3 Apple Apple Orange Orange Pear ...
<p>You can create a list of values from your dataframe select:</p> <pre><code>type1 = list(select['type1'].dropna()) type2 = list(select['type2'].dropna()) type3 = list(select['type3'].dropna()) </code></pre> <p>The use these lists to select the slice of the df</p> <pre><code> df_type1 = df[type1] df_type2 = df[typ...
python|pandas
1
17,025
42,871,395
add rows in pandas dataframe based on date and value
<p>I have a pandas dataframe like the following:</p> <pre><code>id, date, add_days 1, 2017-01-01, 3 2, 2017-03-05, 5 3, 2017-02-27, 3 . . . </code></pre> <p>I want to repeat the ids and increase date by given add_days:</p> <pre><code>id, date, add_days 1, 2017-01-01, 3 1, 2017-01-02, 3 1, 2017-01-03, 3 2, 2017-03-05...
<p>You can use <a href="http://pandas.pydata.org/pandas-docs/stable/generated/pandas.melt.html" rel="nofollow noreferrer"><code>melt</code></a> with <code>groupby</code> and <a href="http://pandas.pydata.org/pandas-docs/stable/generated/pandas.DataFrame.resample.html" rel="nofollow noreferrer"><code>resample</code></a>...
python|pandas
3
17,026
42,588,416
Using for loop to build DataFrame
<p>I am using the following code to make a DataFrame which contains closing prices of two symbols with column heading as their symbol name respectively.</p> <pre><code>symbols=['KEL','PAEL'] start_date = '2016-05-01' end_date = '2017-03-02' allocation = 50000 def data(symbol): dates=pd.date_range(start_date,end_...
<p>your problem is very similar to this one : <a href="https://stackoverflow.com/a/42591863/3027854">https://stackoverflow.com/a/42591863/3027854</a></p> <p>The mistake you are making as pointed out by @A-Za-z is <code>df=data(symbol)</code> you are overwriting your dataframe so you loose the data for <code>KEL</code>...
python-3.x|pandas
0
17,027
42,798,804
Merce csv files (from a folder) into one, adding columns with different names using Python
<p>I need to merge several CSV files that are located in a folder into a single one.</p> <p>My original data is like this</p> <p><strong>y_1980.csv:</strong></p> <pre><code> country y_1980 0 afg 196 1 ago 125 2 alb 23 3 . . . . . </code></pre> <p><str...
<p>Pandas makes this quite easy. With a loop and merge you can simply do:</p> <p><strong>Code:</strong></p> <pre><code>import pandas as pd files = ['file1', 'file2'] dfs = None for filename in files: df = pd.read_csv(filename, sep='\s+') if dfs is None: dfs = df else: dfs = dfs.merge(df,...
python|csv|pandas|merge
1
17,028
26,944,918
Python: looking for duplicates in list
<p>I have a list of floats, and I want to know how many duplicates are in it.</p> <p>I have tried with this:</p> <pre><code>p = t_gw.p(sma, m1, m2) #p is a 1d numpy array p_list = list(p) dup = set([x for x in p_list if p_list.count(x) &gt; 1]) print dup </code></pre> <p>I have also tried to use collections.co...
<p>Your numpy-array is two-dimensional. So <code>list(p)</code> does not do, what you expect. Use <code>list(p.flat)</code> instead.</p> <p>Or (mis)use numpy's histogram function:</p> <pre><code>cnt, bins = numpy.histogram(p, bins=sorted(set(p.flat))+[float('inf')]) dup = bins[cnt&gt;1] </code></pre>
python|numpy|count|hashable
2
17,029
25,247,315
Scipy output of Kolmogorov–Smirnov test
<p>I have made a call for the KS-test function on python to compare two features: size and duration, and I am kind of lost in explaining the output. Here is my code:</p> <pre><code>from scipy.stats import ks_2samp import csv ds1=getColumn(&quot;TraceBG.csv&quot;,5) ds2=getColumn(&quot;TraceFG.csv&quot;,5) ds11=getColu...
<p>The KS test is associated with a null hypothesis: The groups being compared are sampled from the same population.</p> <p>If the null hypothesis is true, the p value is the probability of observing the particular deviation between the sampled groups.</p> <p>For size, the observed level of deviation is to be expecte...
python|numpy
1
17,030
30,387,887
Repeat pandas dataframe over time
<pre><code>import pandas as pd import pdb, random dates = pd.date_range('1950-01-01', '1950-12-31', freq='D') data = [int(1000*random.random()) for i in xrange(len(dates))] cum_data = pd.Series(data, index=dates) </code></pre> <p>How do I repeat this dataframe over the next 10 years?</p>
<p>This should work.</p> <p>For each year in the list of years you provide, I extend the data by the same initial dataset unless it is a leap year.</p> <p>On a leap year, I insert the entry defined by <code>fill_leap</code> at the 60th day (31 days in January + 29 days in February on a leap year)</p> <pre><code>impo...
python|pandas
2
17,031
26,832,866
Finding what elements are in a 2-D array
<p>I'm trying to find what elements are in a 2-D array, such as something along the lines below:</p> <pre><code>import numpy as np a = np.array([[1,0,0],[1,3,0],[2,7,4]]) print find_element(a) [0,1,2,3,4,7] </code></pre> <p>Is there a function that would do this for me?</p>
<p>You could use <a href="http://docs.scipy.org/doc/numpy/reference/generated/numpy.unique.html" rel="nofollow"><code>np.unique</code></a>:</p> <pre><code>&gt;&gt;&gt; a = np.array([[1,0,0],[1,3,0],[2,7,4]]) &gt;&gt;&gt; np.unique(a) array([0, 1, 2, 3, 4, 7]) </code></pre>
python|numpy
4
17,032
39,392,572
numpy vectorization of interdependent arrays
<p>I need to populate two interdependent arrays simultaneously, based on their previous element, like so:</p> <pre><code>import numpy as np a = np.zeros(100) b = np.zeros(100) c = np.random.random(100) for num in range(1, len(a)): a[num] = b[num-1] + c[num] b[num] = b[num-1] + a[num] </code></pre> <p>Is ther...
<p>As mentioned in @Praveen's post, we can write those expressions for few iterations trying to find the closed form and that would be a triangular matrix of course for <code>c</code>. Then, we just need to add in <em>iteratively-scaled</em> <code>b[0]</code> to get full <code>b</code>. To get <code>a</code>, we simply...
numpy|vectorization
3
17,033
39,203,714
List of Structure subclass returns wrong values when casting to numpy array
<p>I've built a simple Structure subclass with two fields, holding a void pointer to an array, and the array length. However, when I try to create a list of these using input lists of the same length, the value of the returned void pointer is the same as the last array used to create the instance:</p> <pre><code>from ...
<p>The problem is that the numpy array is immediately garbage-collected and the underlying memory freed, resulting in a dangling pointer.</p> <p>The solution is to keep a reference to the underlying <code>buffer</code> object:</p> <pre><code>def __init__(self, seq): array = np.array(seq, dtype=np.float64) sel...
python|arrays|numpy|ctypes
2
17,034
19,522,857
About NumPy array in Python
<p>What is the difference between:</p> <pre><code>import numpy as np A = np.zeros((3,)) </code></pre> <p>and</p> <pre><code>import numpy as np B = np.zeros((1,3)) </code></pre> <p>Thanks for your answer!</p>
<p>The first one creates a 1D <code>numpy.array</code> of zeros:</p> <pre><code>&gt;&gt;&gt; import numpy as np &gt;&gt;&gt; A = np.zeros((3,)) &gt;&gt;&gt; A array([ 0., 0., 0.]) &gt;&gt;&gt; A[0] 0.0 &gt;&gt;&gt; </code></pre> <p>The second creates a 2D <code>numpy.array</code> of 1 row and 3 columns, filled with...
python|arrays|numpy
2
17,035
29,057,130
Python 3.4 Pandas 15.2join function locks pc
<p>I'm loading 6 small csv files with 100krecords, indexed by a 15 alphanumeric text field using a laptop with an i5 and 8Gig of ram. </p> <p>I load the files into memory, which shows 9% used. Then I execute a join statement. </p> <pre><code> Df1.join(df2,df3,df4,df5,df6) # join all tables </code></pre> <p>Yes, i...
<p>This seems like it might be because of a silly problem. pandas should give you an error message, but your tables are too large for it to realize that they're the wrong types. Here's what I think is going on:</p> <p><a href="http://pandas.pydata.org/pandas-docs/version/0.15.2/generated/pandas.DataFrame.join.html#pan...
python|join|memory|pandas
0
17,036
29,337,714
How to run naive Bayes from NLTK with Python Pandas?
<p>I have a csv file with feature (people's names) and label (people's ethnicities). I am able to set up the data frame using Python Pandas, but when I try to link that with NLTK module to run a naive Bayes, I get the following error:</p> <pre><code>Traceback (most recent call last): File "C:\Users\Desktop\file.py",...
<p>This line</p> <pre><code>featuresets = [(feature, label) for index, (feature, label) in frame.iterrows()] </code></pre> <p>is choking nbc.train()</p> <p>featuresets should be of the form [(featureset, label)] , where the featureset variable is a <i>dict</i> (not a str) and label is the known class label for the f...
python-2.7|pandas|classification|nltk|naivebayes
1
17,037
29,237,412
Transforming DataFrame
<p>here' a question on transforming a DataFrame. After creating a DataFrame from a pivot table and applying .unstack(), I came up with a table like this</p> <pre> Amount Color Red Yellow Blue Green Date 2006-01 56 41 15 10 2006-01 51 23 26 36 2006-01 36 54 ...
<p>You can get rid of your top-level columns by just selecting the one top-level column value:</p> <pre><code>df = df["Amount"] </code></pre> <p><code>Color</code> isn't the name of a column - it's the name of the row of column names. If you don't want your column levels to have names, you can set them to <code>None...
python|pandas|dataframe
3
17,038
33,899,369
Ranking order per group in Pandas
<p>Consider a dataframe with three columns: <code>group_ID</code>, <code>item_ID</code> and <code>value</code>. Say we have 10 <code>itemIDs</code> total.</p> <p>I need to rank each <code>item_ID</code> (1 to 10) <strong>within</strong> each <code>group_ID</code> based on <code>value</code>, and then see the mean rank...
<p>There are lots of different arguments you can pass to <a href="http://pandas.pydata.org/pandas-docs/version/0.17.0/generated/pandas.Series.rank.html" rel="noreferrer"><code>rank</code></a>; it looks like you can use <code>rank("dense", ascending=False)</code> to get the results you want, after doing a <code>groupby<...
python|pandas
86
17,039
23,520,952
Python: Change format of np.array or allow tolerance in in1d function
<p>I have two numpy arrays (data files loaded with <code>np.loadtxt</code>). They do not have the same length (or number of rows if you will).</p> <p>I want to create a mask, where I find the values in the smaller array in the larger array. For that I can use <code>np.in1d</code>. However, the precision on the larger ...
<p>You could do it as shown below. If <code>a</code> and <code>b</code> get large, this is going to need lots of memory (on the order of the product of the sizes of a and b). Maybe you could loop over into small-enough chunks of b if that's a problem.</p> <pre><code>import numpy as np def in1d_tol(a,b,tol): d=n...
python|arrays|numpy|intersection
2
17,040
22,906,167
How to set xlim for a plot in python?
<p>why doesn't "xlim" work in the following script:</p> <pre><code>from pylab import * from numpy import * x = linspace (0, 2*pi, 100) y = sin (x) clf () plot (x, y, 'b-', label = "sin (x)") xlim (0, 2*pi) axis ('equal') grid () legend () show () </code></pre>
<p>If you leave out the use of <code>axis('equal')</code> aspect ratio, the limitation of the x-values works for me. Possibly this is different for different Matplotlib backends. I am using Mac Os X default backend.</p>
python|python-2.7|numpy|matplotlib
0
17,041
15,436,716
Best way to represent a training set to split with
<p>A training set is made off a set of samples and a set of labels one for each sample. In my case a sample is a vector while a label is a scalar. To deal with this I use Numpy. Consider this example:</p> <pre><code>samples = np.array([[1,0],[0.2,0.5], [0.3,0.8]]) labels = np.array([1,0,0]) </code></pre> <p>Now I hav...
<p>The mixing of indices with float datatypes makes me uneasy. When you say split the training set, is this completely random? If so I would go with the random permutation vector - I don't think your solution is any faster (even without my data type reservations) because you're still allocating memory when creating you...
python|arrays|numpy|artificial-intelligence
1
17,042
29,507,928
Replace slice of a numpy array with values from another array
<p>Say I've got two numpy arrays which were created this way:</p> <pre><code>zeros = np.zeros((270,270)) ones = np.ones((150,150)) </code></pre> <p>How can I insert <code>ones</code> in <code>zeros</code> at position <code>[60,60]</code>? I want an array that looks like a "square in the square".</p> <p>I've tried th...
<p>This is one way you can replace values in zeros with ones. </p> <pre><code>zeros[60:210,60:210] = ones </code></pre>
python|arrays|numpy|insert
7
17,043
62,258,704
What does 'INFO:tensorflow:Oracle triggered exit' mean with keras tuner?
<p>When I run Keras Tuner search, the code runs for some epochs and then says: 'INFO:tensorflow:Oracle triggered exit'.</p> <p>What does this mean? I am still able to extract best hyperparameters. Is it due to early stopping? I have tried both randomsearch and hyperband.</p>
<p>You can solve this with:</p> <pre><code>tuner = RandomSearch( tune_rnn_model, objective='val_accuracy', seed=SEED, overwrite=True, max_trials=MAX_TRIALS, directory='project') </code></pre> <p>To begin a new search and ignore any prior results, we set <code>overwrite=True</code>. Alternativel...
python|tensorflow|keras|neural-network|keras-tuner
3
17,044
62,446,668
pandas substraction by column value
<p>I have one time series on a pandas dataframe that has a row with the month. I've called df1. Then I get the monthly mean by group_by, I've called df2 the resulting dataframe. Now I would like to subtract the monthly mean of each column without using a loop. This is, the row "month==1" in df2, needs to be subtracted ...
<p>Use <a href="http://pandas.pydata.org/pandas-docs/stable/reference/api/pandas.core.groupby.GroupBy.transform.html" rel="nofollow noreferrer"><code>GroupBy.transform</code></a> for new <code>DataFrame</code> filled by aggregate values, so possible subtract by <a href="http://pandas.pydata.org/pandas-docs/stable/refer...
python|pandas|dataframe|array-broadcasting
1
17,045
62,419,264
Update multiple rows in MySQL with Pandas dataframe
<p>I have worked on a dataframe (previously extracted from a table with SQLAlchemy), and now I want to retrieve the changes updating that table.</p> <p>I have done it in this very unefficient way:</p> <pre><code>engine = sql.create_engine(connect_string) connection = engine.connect() metadata = sql.MetaData() pbp = ...
<p>Consider using proper placeholders instead of manually formatting strings:</p> <pre><code>query_update = sql.text(""" UPDATE playbyplay SET Player_1_Visitor = :Player_1_Visitor_y , Player_2_Visitor = :Player_2_Visitor_y , Player_3_Visitor = :Player_3_Visitor_y , Player_4_Visitor = :Player_...
python|mysql|pandas|sqlalchemy
2
17,046
62,250,027
Historic Stock Prices in Pandas DataFrame
<p>This code is able to get the history of stock prices and the dates are set as the index of the DataFrame.</p> <p>How can I pass the dates (index) to a column and add the ticker out of iteration to another column?</p> <pre><code>import yfinance as yf import pandas as pd tickers = ["GOOG","AMZN"] df2 = pd.DataFram...
<p>You can just do this:</p> <pre><code>for ticker in tickers: tkr = yf.Ticker(ticker) hist = tkr.history(period="1y") df2 = df2.append(hist) df2['Ticker'] = ticker print(df2.reset_index()) Date Open High Low Close Volume Dividends Stock Splits Ticker 0 2019-06-06 ...
python|pandas
1
17,047
62,364,429
Android semantic segmentation post-processing is too slow
<p>I'd really appreciate it if anyone can advise with a task I've been working without success for the last week. I have semantic segmentation model (MobileNetV3 + Lightweight ASPP).Short info: input - 1024x1024, output - same size and 2 classes (bg and vehicle), so my output shape is (1, 1048576, 2). I'm not the mobil...
<p>New Update. Thanks to <a href="https://stackoverflow.com/users/7118084/farmaker">Farmaker</a>, I used a piece of code found in his repo from comment above and now pipeline looks like:</p> <pre class="lang-java prettyprint-override"><code> int channels = 3; int n_classes = 2; int float_byte_size = 4; ...
java|android|tensorflow-lite|opencv4android|semantic-segmentation
1
17,048
62,360,340
mode.save() create only one file
<p>i read the documentation of model.save(). it says that it creates the folder but when i save it it just create a single file. when i use this file to change the code from keras to tensorflow-lite it gives this error:</p> <pre><code>OSError: SavedModel file does not exist at: C:\Users\Munib\New folder\3_layer_model_...
<p>I think the model you saved is a <code>keras model</code> and not 'SavedModel<code>. So, you need to use</code>from_keras_model` as shown below. </p> <p>I tried to simulate the issue by using simple model and was successfully reproduced your issue. check below fro the code.</p> <pre><code>!pip install tf-nightly i...
python|tensorflow|keras|model
2
17,049
51,387,753
how to filter string with regular expression in pandas dataframes
<pre><code>import pandas as pd data = [['Alex',10],['Bob',12],['Clarke',13] ['Adam', 14]] df = pd.DataFrame(data,columns=['Name','Age']) print(df) Name Age 0 Alex 10 1 Bob 12 2 Clarke 13 3 Adam 14 </code></pre> <p>I want to get only Names starting with A . I tried following code mask = d...
<p>Use this:</p> <pre><code>mask = df['Name'].str.startswith("A") </code></pre> <p>For example:</p> <pre><code>In [52]: df Out[52]: Name Age 0 Alex 10 1 Bob 12 2 Clarke 13 3 Adam 14 In [53]: mask = df['Name'].str.startswith("A") In [54]: df[mask] Out[54]: Name Age 0 Alex 10 3 Ada...
python|pandas
4
17,050
51,186,374
Import System Modules in a Conda Environment
<p>I have installed caffe and pytorch0.3 in my system path(environment), and there is a project which only works under caffe and history version of pytorch0.2. To solve that I install pytorch0.2 in a Conda environment and I wonder if there is any way can save me from installing a caffe again in this conda environment. ...
<p>Activate your environment and try the following</p> <pre><code>conda install package-name --offline </code></pre> <p>Also, in case you wish to clone the root or some other environment to a conda environment, you can use -- clone. For instance, when you wish to clone the root - </p> <pre><code>conda create -n pyto...
anaconda|conda|pytorch
1
17,051
51,529,840
Pandas combine duplicate columns that contain strings
<p>I'm having problems (sort of) with combining duplicate columns. It seems to work on older versions of Pandas/Python (not sure what the culprit is here), but not on the latest version. </p> <p>I basically have a dataframe of mixed values with duplicate column names after a concat. The values are either going to be a...
<p>I think you need <code>transpose</code> the dataframe firstly, <code>reset the index</code>, and then <code>rename</code> the duplicate <code>index</code> column values, and lastly use <code>groupby</code>.</p> <pre><code>df_t = df.T.reset_index() df_t["index"] = df_t["index"].str.split(".").str[0] result = df_t.gr...
python|pandas|pandas-groupby
1
17,052
51,171,560
compare two columns having list of strings in pandas
<p>I have a data frame in pandas having two columns where each row is a list of strings, how would it be possible to check if there is word match(es) in these two columns on a unique row(flag column is the desired output)</p> <pre><code>A B flag hello,hi,bye bye, also 1 but, as wel...
<p>You can convert each value to separately words by split and <code>set</code>s and check intersection by <code>&amp;</code>, then convert values to boolean - empty sets are converted to <code>False</code>s and last convert it to <code>int</code>s - <code>Falses</code> are <code>0</code>s and <code>True</code>s are <c...
python|pandas
3
17,053
48,067,854
Trouble understanding tensorflow shuffle_batch enqueue_many=False
<p>I am reading the Tensorflow documentation and the code for the Cifar10 example. This bit is currently racking my brain:</p> <pre><code># Creates batches of 32 images and 32 labels. image_batch, label_batch = tf.train.shuffle_batch( [single_image, single_label], batch_size=32, num_threads=4, capacity=50000,...
<p>The <code>single_image</code> or <code>single_label</code> tensor would usually refer to an operation that retrieves the next value from a queue. To create a batch, it would then for example retrieve the batch size (e.g. 32) of values from those tensors if it wasn't shuffled. In the case where it is shuffled it will...
python|tensorflow
1
17,054
48,591,713
Pearson correlation and nan values
<p>I have two CSV_files with hundreds of columns and I want to calculate Pearson correlation coefficient and p value for every same columns of two CSV_files. The problem is that when there is a missing data "NaN" in one column, it gives me an error. When ".dropna" removes nan value from columns, sometimes the shapes of...
<p>Here is one solution. First calculate the "bad" indices for your 2 numpy arrays. Then mask to ignore those bad indices.</p> <pre><code>x = np.array([5, 1, 6, 9, 10, np.nan, 1, 1, np.nan]) y = np.array([4, 4, 5, np.nan, 6, 2, 1, 8, 1]) bad = ~np.logical_or(np.isnan(x), np.isnan(y)) np.compress(bad, x) # array([ ...
python|arrays|numpy|nan|pearson-correlation
12
17,055
48,777,345
Vectorized random walk in Python with boundaries
<p>I'm trying to simulate a 2-D random walk in python with boundaries (the particle/object will can't cross the boundaries and must return back). However my version is not vectorised and is very slow. How do I implement it without(or minimising) the use of loops.</p> <p>Here is my approach</p> <pre><code>def bound_wa...
<p>Here is one approach that is fast, but not 100% equivalent to your implementation. The difference is that in my implementation, at the boundary the chance of going in one of the directions parallel to the boundary is half that of the directions retreating from the boundary. That is arguably the better model if you t...
python|python-2.7|numpy|scipy
1
17,056
48,653,893
merging two dataframes with same rows and indexes in pandas
<p>I'm trying to merge two pandas dataframes that have common row indexes and common columns 0,1,2 but different column 3, so the resulting dataframe has columns from both:</p> <p>First dataframe:</p> <pre><code>&lt;class 'pandas.core.frame.DataFrame'&gt; RangeIndex: 817 entries, 0 to 816 Data columns (total 3 column...
<p>I would try:</p> <pre><code>pd.merge(df, df2, on=['0', '1']) </code></pre> <p>maybe</p> <pre><code>pd.merge(df, df2, on=[0,1] </code></pre>
python|pandas|dataframe|merge
3
17,057
48,455,129
Pip is correctly installing libraries to the proper directory, but I cannot import those packages properly in program
<p>I installed (with pip) MatPlotLib and Pandas, and they are both not working properly in programs. Here is the strange thing... When I type the following into the interactive environment of IDLE import pandas as pd</p> <p>pd.Series([1, 2, 3, 4, 5])</p> <p>I get this as output: (indicating that it works properly)</...
<p>SOLVED. So when I did:</p> <p>import sys</p> <p>sys.path.append(path to pandas library) </p> <p>it worked! so now I can fully use pandas. I guess I will just have to do this anytime I download a new library and it doesn't work. Thank you for all the help</p>
python|python-3.x|pandas|python-import
0
17,058
70,744,865
Make a Column Based on Pandas Timeseries Minutes OHLCV Dataframe Where New Column Fills with Corresponding Day's First Value from a Specific Column
<p>My initial <strong>Pandas Dataframe</strong> is:</p> <pre><code> Open High Low Close Volume 2021-11-19 09:30:00-05:00 16549.50 16559.25 16516.25 16530.25 20198.0 2021-11-19 09:35:00-05:00 16530.50 16562.00 16525.50 16556.50 11274.0 .....
<p>Here is my <strong>solution</strong>. May be there are better solutions than this solution.</p> <pre><code>#Here df is our initial ohlcv dataframe #Groupby date and timestamp df1 = pd.DataFrame(df.groupby([df.index.date, df.index], axis=0)['Open'].first()) #Make Timestamp as column df1.reset_index(level=1, inplace...
python|pandas|dataframe|numpy|time-series
0
17,059
70,764,910
convert dataframe in python to json or dictionary type object
<p>I have data frame like</p> <pre><code>data = pd.DataFrame({'col1' :['A','A','A','B','B','C','C','C'], 'col2': ['13','15','17','11','15','12','21','23'], 'col3' : [3,5,8,7,2,5,1,3]},columns= ['col1', 'col2', 'col3']) print(data) col1 col2 col3 0 A 13 3 1 A 15 5 2 A 17 8 3 B 11 ...
<p>You'll have to bite the bullet and use a lambda with a groupby statement.</p> <p>Note lambda's aren't very efficient and this doesn't really seem like a logical / proper json structure.</p> <p>but to answer your question.</p> <pre><code>data.groupby('col1')\ .apply(lambda x : {'col2' : dict(zip(x['col2'],x[...
python|pandas
0
17,060
70,884,452
Formatting column in pandas to decimal places using table.style based on value
<p>I am trying to format a column in a dataframe using style. So far I successfully used the styling for a fixed number of decimals:</p> <p>mytable.style.format('{:,.2f}', pd.IndexSlice[:, ['Price']])</p> <p>but I need to expand this to formatting based on value as this:</p> <ul> <li>if value is &gt;=1000, then format ...
<p>Building upon @Code_beginner's answer – the callable should return formatted string as output:</p> <pre><code>def my_format(val): if val &gt;= 1000: return f&quot;{val:,.0f}&quot; if val &gt;= 1: return f&quot;{val:,.2f}&quot; return f&quot;{val:,.5f}&quot; mytable.style.format({'Price':...
pandas|dataframe
2
17,061
70,755,256
How to add elements of an array to a numpy array to represent coordinates?
<p>Hey people of Stackoverflow, I want to achieve the following procedure. I have these 2 arrays:</p> <pre><code>a = np.array([175, 370]) b = [array([175, 176, 176, 176]), array([371, 369, 370, 371])] </code></pre> <p>the elements of b array should represent coordinates of some points</p> <pre><code>b = [array[x1, x2, ...
<p>You can try the <a href="https://docs.python.org/3.9/library/functions.html#zip" rel="nofollow noreferrer">zip</a> function to pair two arrays in <code>b</code> and append it to <code>a</code></p> <pre><code>import numpy as np a = [175, 370] b = [np.array([175, 176, 176, 176]), np.array([371, 369, 370, 371])] a = ...
python|arrays|numpy
0
17,062
70,839,909
Merge pandas dataframes and transform in SQL
<p>I have few dataframes:</p> <p><strong>DF_1</strong></p> <pre><code>| | TYPE | SYMBOL | DESCRIPTION | OPOL | FIRST_INTEREST_DATE | |----|------|--------|-------------|------|---------------------| | 0 | BOND | 1 | FIRST | 10 | 20220531 | | 1 | BOND | 2 | SECOND | 20 | ...
<p>you need to upload code or show a sample of your data.</p> <p>Anyways here's a solution based on a limited information that you've provided:</p> <p><code>df_3 = pd.concat([df_1, df_2], axis=0)</code></p> <p>then you can use</p> <pre class="lang-py prettyprint-override"><code>from sqlalchemy import create_engine engi...
python|sql|pandas|dataframe
1
17,063
70,973,508
Calculate amount of rows satisfy condition in multiple columns
<pre><code>import pandas as pd # initialize list of lists data = [['tom', 'Y','Y','N'], ['nick', 'N','N','N'], ['juli', 'N','Y','N'], ['Luc', 'Y','Y','N'], ['Adg', 'Y','N','N'], ['Flav', 'N','Y','N'], ['Alf', 'Y','Y','N'], ['Jut', 'Y','N','N'], ['Uan', 'Y','Y','Y']] # Create the pandas DataFrame df = ...
<p>just group by them:</p> <pre><code>df = df.groupby(['Tipo_1', 'Tipo_2','Tipo_3']).count().reset_index() print(df) </code></pre> <p>output:</p> <pre><code>&gt;&gt;&gt; Tipo_1 Tipo_2 Tipo_3 Name 0 N N N 1 1 N Y N 2 2 Y N N 2 3 Y Y N 3 4 ...
python|pandas|plot
4
17,064
51,596,059
python cx_oracle hang when storing as DataFrame?
<p>I'm trying to store the results of an Oracle SQL query into a dataframe and the execution hangs infinitely. But, when I print the query it comes out instantly. What is causing the error when saving this as a DataFrame?</p> <pre><code>import cx_Oracle import pandas as pd dsn_tns = cx_Oracle.makedsn('HOST', 'PORT', s...
<p>Pandas's <a href="https://pandas.pydata.org/pandas-docs/stable/generated/pandas.read_sql.html" rel="nofollow noreferrer"><code>read_sql</code></a> requires a connection object for its <em>con</em> argument not the result of a cursor's <code>execute</code>. Also, consider using <a href="http://docs.sqlalchemy.org/en/...
python|oracle|pandas
0
17,065
51,618,724
Reshaping dataframe from vertical to horizontal in python
<p>I have data samples from a database which I want to reshape from vertical to horizontal in python for further data analysis. The dataframe looks like this:</p> <pre><code>ID measured_at weight aa 2017-11-04 78.1 bb 2018-04-08 74.2 bb 2018-04-16 73.2 bb 2018-04-28 72.1 cc 2018-03-02 90.2 cc 2018-0...
<p>You can use <code>GroupBy</code> with <code>list</code> a couple of times, followed by <code>pd.concat</code>.</p> <p>I leave renaming columns and elevating index to column as an exercise.</p> <pre><code>g = df.groupby('ID') df_dates = pd.DataFrame(g['measured_at'].apply(list).values.tolist(), index=g.groups) df_w...
python|pandas|dataframe|data-analysis
0
17,066
51,831,526
pandas merge multiple dataframes
<p>For example: I have multiple dataframes. Each data frame has columns: variable_code, variable_description, year. </p> <p>df1:</p> <pre><code>variable_code, variable_description N1, Number of returns N2, Number of Exemptions </code></pre> <p>df2:</p> <pre><code>variable_code, variable_description N1, ...
<p>First, concatenate df1, df2, by using </p> <pre><code> final_df = pd.concat([df1,df2]). </code></pre> <p>Then we can convert columns variable_code, variable_name into dictionary. variable_code as keys, variable_name as values by using </p> <pre><code> d = dict(zip(final_df['variable_code'], final_df['variable_na...
pandas|dataframe
0
17,067
41,719,812
Error while calling eval() on Tensor variable in keras
<p>I am using keras and using a layer output for some modifications. Before, using the output ( a tensor variable ) I am converting it to numpy array and thus calling eval() on it, as below:</p> <pre><code>def convert_output(orig_output): conv_output = invoke_modifications(orig_output.eval(), 8) </code></pre> <p>...
<p>You don't need to call eval() at all, your conversion_method should be done using symbolic functions (the ones from keras.backend) and should be differentiable.</p> <p>It won't work otherwise, the network won't be able to be trained with Keras/Theano.</p>
python|numpy|theano|keras|keras-layer
1
17,068
47,916,065
Summing edges in pandas
<p>Have a dataframe that represents a network with directed edges between nodes A, B and C. I want to sum the flows between nodes to measure the strength of connections between nodes. Below is the df I have..</p> <pre><code>To From Flow A A 1 A B 4 A C 2 B A 5 B B 2 B C 6 C A 3 C B 5 C...
<p>Here is one way to solve this.</p> <pre><code>df["Node 1"] = df[["To","From"]].min(axis=1) df["Node 2"] = df[["To","From"]].max(axis=1) result = df.groupby(["Node 1","Node 2"], as_index=False)["Flow"].sum() </code></pre> <p>The first two lines just create a consistent order for the groups. The third line just adds...
python|pandas
1
17,069
48,967,207
Convert Time with AM/PM and +UCT (Python)
<p>I'm trying to convert <code>'11/09/2011 11:33:00 PM +0000'</code> (object) into a datetime variable using Pandas' <code>pd.to_datetime()</code>:</p> <pre><code>df['Datetime'] = pd.to_datetime(df['Datetime'],format = '%d/%m/%Y %I:%M:%S %p', utc=True) </code></pre> <p>however getting a error back:</p> <blockquote>...
<p>If the <code>+0000</code> is fixed you can add it to the format string like:</p> <h3>Code:</h3> <pre><code>pd.to_datetime(df['Datetime'], format='%d/%m/%Y %I:%M:%S %p +0000', utc=True) </code></pre> <h3>Test Code:</h3> <pre><code> df = pd.DataFrame([ '11/09/2011 11:33:00 PM +0000', '11/09/2011 11:33:00 P...
python|pandas|datetime
1
17,070
49,140,589
Filling Pandas columns with lists of unequal lengths
<p>I am having trouble filling Pandas dataframes with values from lists of unequal lengths.</p> <p><code>nx_lists_into_df</code> is a list of numpy arrays.</p> <p>I get the following error: </p> <blockquote> <p>ValueError: Length of values does not match length of index</p> </blockquote> <p>The code is below:</p>...
<p>We'll leverage <code>pd.Series</code> to attach an appropriate index and will allow us to use the <code>pd.DataFrame</code> constructor without complaining of unequal lengths.</p> <pre><code>df1, df2 = ( pd.DataFrame(dict(zip(df_cols, map(pd.Series, d)))) for d in nx_lists_into_df ) </code></pre> <hr> <pr...
python|python-3.x|pandas|numpy
1
17,071
58,998,132
Pandas dataframe conditional column based on multiple conditions only working on first condition?
<p>I have a data frame that looks something like this: (there are about 100 more columns irrelevant to my conditional column calculation)</p> <pre><code>col1 col2 col3 a NaN NaN b NaN NaN NaN a NaN NaN b NaN NaN NaN a NaN NaN b </code></pre> ...
<p>Let us try <code>bfill</code></p> <pre><code>df['col4']=df.bfill(1).iloc[:,0] df Out[107]: col1 col2 col3 col4 0 a NaN NaN a 1 b NaN NaN b 2 NaN a NaN a 3 NaN b NaN b 4 NaN NaN a a 5 NaN NaN b b </code></pre>
python|pandas|dataframe
4
17,072
58,960,685
Joint distribution from pandas dataframe?
<p>I have a pandas dataframe that has two columns <code>A</code> and <code>B</code>, both of which contain numbers. I want to create a joint distribution from these columns, ie. I want to bin both columns (where each bin of <code>A</code> contains every bin of <code>B</code>, and so on for a joint distribution) and ass...
<p>You can cut <code>A</code> and <code>B</code> into their own bins, then apply a <code>groupby</code> to count them:</p> <pre><code># Some sample input data np.random.seed(42) df = pd.DataFrame(np.random.randint(1, 1000, (1000, 2)), columns=['A', 'B']) # Assign values in columns A &amp; B into 5 bins each # The bin...
pandas
0
17,073
58,956,597
Elegant mapping of GroupBy into dataframe
<p>In order to best decide how to handle missing data of each feature in a weather dataset, I want to get the length of the longest block of NaNs for each feature and each weather station, the latter being denoted by 'id'. Although the following accomplishes this, I am aware of the code's awkwardness. What would be an ...
<p>How about applying your function to the groupby using <code>agg</code>:</p> <pre><code>result = weather_df.groupby('id').agg({ 'temperature': max_repeated_nans, 'wind_speed': max_repeated_nans }) </code></pre>
python|pandas|dataframe|pandas-groupby|missing-data
1
17,074
70,132,755
Pandas fill NaN in columns based on some conditions
<p>I have a pandas dataframe which consists of weekly values for a given id, looking something like :</p> <div class="s-table-container"> <table class="s-table"> <thead> <tr> <th>id</th> <th>val_week1</th> <th>val_week2</th> <th>val_week3</th> <th>val_week4</th> <th>val_week5</th> </tr> </thead> <tbody> <tr> <td>1</td>...
<pre class="lang-py prettyprint-override"><code>temp = df.filter(like='val') temp = temp.mask(temp.cumsum(1).eq(0), np.nan) df.assign(**temp) id val_week1 val_week2 val_week3 val_week4 val_week5 0 1 NaN NaN 3.0 2.0 0 1 2 1.0 0.0 0.0 2.0 ...
python|pandas|dataframe
3
17,075
70,152,993
Python: Intersect 2 lists mucisians with excluding the ratings from the list with basic tools
<p>Trying to intersect 2 lists of musicians with their rating in each specific list. Want to find similarities: whose musicians who are in both lists, but to do so I have to exclude the rating. First I tried to write the code to find the same musicians for lists without ratings, however got empty output []. May be some...
<p>Modify your intersection function as such</p> <pre><code>def intersect(list_of_names1, list_of_names2): list1 = [&quot; &quot;.join(line[:-1]) for line in list_of_names1] list2 = [&quot; &quot;.join(line[:-1]) for line in list_of_names2] list3 = list(set(list1) &amp; set(list2)) return list3 p...
python-3.x|numpy|intersect
1
17,076
56,296,468
Get variable scope in tensorflow 2.0
<p>I am using the new version of tensorflow (2.0.0 alpha) and I dont know why it does not allow me to run:</p> <pre><code>tf.get_variable_scope().reuse_variables() </code></pre> <p>I am getting the following error:</p> <pre><code>AttributeError: module 'tensorflow' has no attribute 'get_variable_scope' </code></pre...
<p>From comments</p> <blockquote> <p>No, <code>variable_scope</code> and <code>get_variable</code> have been removed from <code>TensorFlow 2.x</code> (although you can still find them under <a href="https://www.tensorflow.org/api_docs/python/tf/compat/v1" rel="nofollow noreferrer">tf.compat.v1</a>). In <code>TF 2.x</co...
python|tensorflow|variables|scope
1
17,077
55,870,089
How to create a new column with values after groupby & rollingsum?
<p>I am trying to create a new column in an existing <code>df</code>. The values of the new column are created by a combination of the groupby and rolling sum. How do I do this?</p> <p>I've tried two approaches both resulting in either NaN values or 'incompatible index of the inserted column with frame index'</p> <p>...
<p>To ensure alignment on the original (non-duplicated) index:</p> <pre><code>df.groupby('HomeTeam', as_index=False)['FTHP'].rolling(4).sum().reset_index(0, drop=True) </code></pre> <hr> <p>With a <code>df</code>:</p> <pre><code> HomeTeam FTHP A a 0 B b 1 C b 2 D a 3 E ...
python|pandas|group-by|multiple-columns|rolling-sum
0
17,078
64,983,913
How to vectorize pandas operation
<p>I have a dataset of house sales with timestamped Periods(per quarter). I want to adjust the price according to the house pricing index change per region. I have a separate dataframe with 3 columns, the Quarter, the Region and the % change in price. I am currently achieving this by iterating over both dataframes. Is ...
<p>Use <a href="http://pandas.pydata.org/pandas-docs/stable/reference/api/pandas.DataFrame.merge.html" rel="nofollow noreferrer"><code>DataFrame.merge</code></a> with <code>left_on</code> and <code>right_on</code>, then get all 4 column in output:</p> <pre><code>df = houses_df.merge(HPindex_df, le...
python|pandas|dataframe
2
17,079
64,646,490
Calculate similarity between rows of a dataframe (count values in common)
<p>I want to calculate similarity between the rows of my dataframe. I have some columns with informations about some people. One row is one person. It looks like that :</p> <pre><code> print(df) id name firstname email town age 0 1 martin pierre truc@machin.com Paris...
<p>You can use <a href="https://docs.scipy.org/doc/scipy/reference/generated/scipy.spatial.distance.pdist.html" rel="noreferrer"><code>scipy.spatial.distance.pdist</code></a> with a custom distance function</p> <pre><code>from scipy.spatial.distance import pdist, squareform pd.DataFrame(1 - squareform(pdist(df.set_inde...
python|python-3.x|pandas|dataframe
6
17,080
64,718,049
Multiindex Re-indexing - custom sort
<p>How do you reindex a multiindex? I have 2 indexes, where the first one requires a manual sort, and the 2nd index is by descending order.</p> <p>In the below, index1 needs to be custom sorted by: Cat-B, Cat-A, Cat-C; index2 by descending order.</p> <pre><code>Index1 Index2 Cat-A Apple Orange Banana Cat...
<p>You need <code>pd.CategoricalIndex</code> however since this orders in one direction we will manually need to invert the order.</p> <pre><code>cati = pd.CategoricalIndex( df.index.get_level_values(0).unique(), [&quot;Cat-C&quot;, &quot;Cat-A&quot;, &quot;Cat-B&quot;], ordered=True ) #note 'Cat-C' is first but wi...
python|pandas|dataframe|sorting
0
17,081
64,627,576
Pass weights into CrossEntropyLoss in correct order
<p>I'm trying to use weight in <code>torch.nn.CrossEntropyLoss</code>,</p> <p>but I'm not sure which order should I put</p> <p>e.g.</p> <pre><code> weight = torch.tensor([1.0, 52337/34649, 52337/11066]).to(device) criterion = nn.CrossEntropyLoss(weight=weight) </code></pre> <p>My class0 has 52337 examples ...
<p>after you define your criterion as above you will have to call it like:</p> <pre><code>loss = criterion (inputs, targets) </code></pre> <p>the targets will be encoded 0 .. C-1 where C is number of classes (0,1 or 2 in your case). So your weights order and number should correspond to your targets.</p>
python|pytorch
0
17,082
64,789,762
add rows to pandas dataframe based on days in week
<p>So I'm fairly new to pandas and I run into this problem that I'm not able to fix.</p> <p>I have the following dataframe:</p> <pre><code>import pandas as pd df = pd.DataFrame({ 'Day': ['2018-12-31', '2019-01-07'], 'Product_Finished': [1000, 2000], 'Product_Tested': [50, 10]}) df['Day'] = pd.to_datetime(d...
<p>You can achieve this by first creating a new DataFrame that contains the desired date range using pandas.date_range.</p> <p>Step 2, use pandas.merge_asof specifying to get the last value.</p>
python|pandas|dataframe
0
17,083
40,251,658
How to fill missing DataFrame value by copying a value from another column
<p>Mick Jagger's last name is missing in a data. Only the fist name and the band's name were specified </p> <pre><code>import pandas as pd df = pd.DataFrame({ 'fist': ['John', 'Mick'], 'last':['Lennon', ''], 'band':['Beatles', 'Rolling Stones'] }) </code></pre> <p>I can easi...
<p>a missing data would normally be a NaN values, not a string. So in the case that you have NaNs instead of '', you can actually pass another <code>column</code> to the <code>fillna()</code> method:</p> <pre><code>import pandas as pd import numpy as np df = pd.DataFrame({ 'fist': ['John', 'Mick'], ...
python|pandas|dataframe
2
17,084
40,101,130
how do I calculate a rolling idxmax
<p>consider the <code>pd.Series</code> <code>s</code></p> <pre><code>import pandas as pd import numpy as np np.random.seed([3,1415]) s = pd.Series(np.random.randint(0, 10, 10), list('abcdefghij')) s a 0 b 2 c 7 d 3 e 8 f 7 g 0 h 6 i 8 j 6 dtype: int64 </code></pre> <p>I want to get the...
<p>There is no simple way to do that, because the argument that is passed to the rolling-applied function is a plain numpy array, not a pandas Series, so it doesn't know about the index. Moreover, the rolling functions must return a float result, so they can't directly return the index values if they're not floats.</p...
python|pandas|numpy|dataframe|series
14
17,085
40,900,034
Python - Extract matrix a into matrix b
<p>I would like to extract matrix A (4,4) into matrix b (7,4).</p> <p>Below matrix :</p> <pre><code>&gt;&gt;&gt; Matrix_A = numpy.array([[[10,1.5,-3.8,8.0],[20,10.2,5.2,6.7],[30,0.5,-6.2,-7.1],[40,-0.7,-0.6,-0.5]]]) &gt;&gt;&gt; Matrix_A array([[[ 10. , 1.5, -3.8, 8. ], [ 20. , 10.2, 5.2, 6.7], ...
<p>First off all since your arrays ar 3d arrays you should squeeze the size to convert them to 2d arrays for the ease of computation.</p> <pre><code>In [27]: Matrix_A = np.squeeze(Matrix_A) In [28]: Matrix_B = np.squeeze(Matrix_B) </code></pre> <p>Then you can use <code>np.in1d</code> to find the indices of the comm...
python|numpy|matrix|compare
1
17,086
41,048,270
Next day or next row index in Pandas Data frame
<p>I am trying to find a way to get the next day (next row in this case) in a Pandas dataframe. I thought this would be easy to find but Im struggling. </p> <p><strong>Starting Data</strong>:</p> <pre><code> ts = pd.DataFrame(np.random.randn(1000), index=pd.date_range('1/1/2000', periods=1000)) ts.columns = ['Va...
<p>I managed to work this one out so sharing the answer:</p> <pre><code> ts.loc[tsSig.index + pd.DateOffset(days=1)] tsSig['Val_Dayplus1'] = ts['Val'].ix[tsSig.index + pd.DateOffset(days=1)].values tsSig Val Week Val_Dayplus1 2000-02-15 1.551125 7 -0.102154 2000-02-24 1.525402 8 -...
python|pandas
2
17,087
41,023,733
Use Google Cloud Machine Learning service to predict with a locally retrained Inception model
<p>I have locally retrained the Inception model using the <a href="https://github.com/tensorflow/tensorflow/blob/master/tensorflow/examples/image_retraining/retrain.py" rel="noreferrer">retrain.py</a> file from Google Code Lab TensorFlow for Poets and want to use Google Cloud machine Learning service to make prediction...
<p>This <a href="https://cloud.google.com/blog/big-data/2016/12/how-to-train-and-classify-images-using-google-cloud-machine-learning-and-cloud-dataflow" rel="nofollow noreferrer">posting</a> yesterday by Google's <a href="https://stackoverflow.com/users/6915676/slaven-bilac">Slaven Bilac</a> appears to be the answer.</...
tensorflow|google-cloud-ml
3
17,088
53,946,436
multiple index to columns in pandas
<p>I have a dataframe in pandas:</p> <pre><code>(index) amount 0.0 0 73.74770979 0.0 1 34.36146516000001 1.0 0 25.759792399999995 1.0 1 117.37044276999995 </code></pre> <p>I would like to have a DataFrame like this:</p> <pre><code>index amount_0, amount_1 0.0 73.... 34... 1.0 25.... 117... </co...
<p>I believe <code>df.unstack(level=-1)</code> should do the trick.</p>
python|python-3.x|pandas|dataframe
2
17,089
66,135,132
Apply dictionary to dataframe column containing strings
<p>I am trying to apply a dictionary to a dataframe column that contains rows of strings, eg:</p> <p>0 [some, text, etc...</p> <p>1 [other, text...</p> <p>And have used the code below that works on a single string but not for a dataframe:</p> <pre><code>dict = pd.read_csv(r'C:\Users\dictionary.csv', header=...
<p>This error can come out if you have NaN value in your Data Series</p>
pandas|dictionary|normalization
0
17,090
66,240,259
Prediction with ANN are most of the time false
<p>my probleme is that my ANN predicts only about 2 times out of 10 the right digit but when the ANN was fitted it tells me about 98% accuarcy. I'm quit a starter with ANNS and I don't know if I'm missing something abvious or why it is like it is. For testing I use a tabel with 81 digits (allways in a row from 1 to 9)....
<p>One thing you can do is add validation_data in model.fit like this :</p> <pre><code> model.fit(x_train, y_train, epochs=50, validation_data=(x_test,y_test)) </code></pre> <p>You can also add some 2Dconv layers with pooling before the flatten one. Or add more neurones.</p> <p>Let me know if it helps.</p>
python|tensorflow|keras|mnist|digits
0
17,091
66,235,771
Need a more efficient way of creating a dictionary of dataframes from a single large dataframe
<p>So the issues is that I have a large dataframe (some millions of rows) and I need to split it into separate dfs based on a value of metric (which can have several thousand unique values in the df) and then put all individual dfs into a dictionary.</p> <p>The data looks like this:</p> <pre><code>&gt;&gt;&gt; df.sampl...
<pre class="lang-py prettyprint-override"><code>df_grouped = df.groupby('metric') mhi_dict = {} for key in df_grouped.groups: group = df_grouped.get_group(key) mhi_dict[key] = group </code></pre>
python|pandas
1
17,092
52,852,742
how to read tensorflow confusion matrix rows and columns
<p>For my 2 classes (<code>1 = [0, 1]</code> and <code>0 = [1, 0]</code>) CNN model I use <code>tf.confusion_matrix</code> to finding a confusion matrix for the model. one of my results is like below for validation set:</p> <pre><code>[ [1800 17] [283 600] ] </code></pre> <p>after doing some search I see more th...
<p>The truth is behind the code ;) <a href="https://github.com/tensorflow/tensorflow/blob/master/tensorflow/python/ops/confusion_matrix.py" rel="nofollow noreferrer">https://github.com/tensorflow/tensorflow/blob/master/tensorflow/python/ops/confusion_matrix.py</a></p> <blockquote> <p>Class labels are expected to s...
python-3.x|tensorflow
1
17,093
52,631,632
Regex: How to capture words with spaces/hyphens excluding numbers?
<p>I have a dataset that looks like this:</p> <pre><code>Column1 ------- abcd - efghi 1234 aasdas - asdas 54321 asda-asd 2344 aasdas(asd) 5234 </code></pre> <p>I want to be able to pull everything out that will exclude a number so it will look like this:</p> <pre><code>Column2 ------- abcd - efghi aasdas - asdas asd...
<p>Like using <code>replace</code> </p> <pre><code>df.Column1.str.replace('\d+','') Out[775]: 0 abcd-efghi 1 aasdas-asdas 2 asda-asd 3 aasdas(asd) Name: Column1, dtype: object #df.Column1=df.Column1.str.replace('\d+','') </code></pre>
python|regex|string|pandas
2
17,094
52,873,451
Pandas: difference between groups
<p>Hi I have a panda df that looks like the following (not real data)</p> <pre><code>index datedjourney sequence values 1 1 1 120 2 1 1 100 3 1 2 75 4 1 3 50 5 1 3 30 6 ...
<p>Create a "diff_between_groups" column which is the difference between "values" and "values" shifted one row.</p> <p>Make a boolean mask to find the rows where "datedjourney" is the same and "sequence" is different from the row above.</p> <p>Use pandas Series where function to replace "diff_between_groups" values u...
python|pandas
2
17,095
46,625,632
tensorflow reduce_mean vs numpy mean
<p>As far as I understand tensorflow reduce_mean and numpy mean should return identical values, but below example returns different values:</p> <pre><code>import numpy as np import tensorflow as tf t_1 = tf.constant([1,3,4,5]) t_2 = tf.constant([7,8,9,0]) list_t = [t_1, t_2] reduced_t_list = tf.reduce_mean(list_t) se...
<p>From <a href="https://www.tensorflow.org/versions/r0.12/api_docs/python/constant_op/constant_value_tensors#constant" rel="nofollow noreferrer"><code>tf.constant</code> docs</a>:</p> <pre><code>If the argument dtype is not specified, then the type is inferred from the type of value. </code></pre> <p>The <code>dtype...
numpy|tensorflow|mean
1
17,096
68,926,632
Create Repeating N Rows at Interval N Pandas DF
<p>i have a df1 with shape 15,1 but I need to create a new df2 of shape 270,1 with repeating rows from each row of the rows in df1 at intervals of 18 rows 15 times (18 * 15 = 270). The df1 looks like this:</p> <pre><code> Sites 0 TULE 1 DRY LAKE I 2 PENASCAL I 3 EL CABO...
<p>I FINALLY found the answer: convert the dataframe to a series and use repeat in the form: my_series.repeat(N) and then convert back the series to a df.</p>
pandas|dataframe|rows|repeat
0
17,097
44,571,370
Filtering a dataframe from information in another dataframe using Pandas
<p>I have a dataframe that is below.</p> <pre><code>df = pd.DataFrame(columns=['Chromosome', 'Start','End'], data=[ ['chr1', 2000, 3000], ['chr1', 500, 1500], ['chr3', 3000, 4000], ['chr5', 4000, 5000], ['chr17', 9000, 10000], ['chr19', 1500, 2500]...
<p>You can try this:</p> <pre><code>df.merge(probes, left_on='Chromosome', right_on='Chrom').query('Start &lt; Position &lt; End') </code></pre> <p>Output:</p> <pre><code> Chromosome Start End Probe Chrom Position 0 chr1 2000 3000 CG999 chr1 2500 2 chr19 1500 2500 CG000 chr19 2...
python|pandas|filter
5
17,098
60,894,305
Memory problem using python pandas to join stock DataFrames in loop
<p>I am trying to join a lot of dataframes in order to do the correlation matrix in pandas. So, it seems that I have to keep on adding columns on the right hand, with the "Date" as the index. But, when I try to do this function with just 50 dataframes, it ends with the memory error.</p> <p>Is there anyone knows what i...
<p><code>Pandas</code> is not designed for such dynamic concatenations. You could just append things into a list, and convert that list into a DataFrame. Like so:</p> <pre><code>join=[] for key, value in stock_dic.items(): join.append({'Date':value} ) df_join=pd.DataFrame(join) </code></pre>
python|mysql|pandas
1
17,099
60,912,882
Is there a way to print a pandas data frame as copyable code?
<p>Every now and then I need to debug code and arrive to a point where I find a valid "oracle" or "fixture" for a new test case and then need to print the data frame to the console and turn it into a data frame initialization code. Is there a way to print a data frame into copyable code? I know I could just dump it in ...
<p>If what you seek is representation of the DataFrame that could be use as valid Python code as suggest you use <a href="https://pandas.pydata.org/pandas-docs/stable/reference/api/pandas.DataFrame.to_dict.html" rel="nofollow noreferrer">to_dict</a>, as follows:</p> <pre><code>import pandas as pd df = pd.DataFrame(da...
python|pandas
2