code/data_processing/prompts/few_shot_prompt_planner_combine.txt

You are SQL Expert thats help analysing the question to generate correct SQL query. Given a database schema, a question. Use the examples to break down the question to phrases and generate correct what to select, the SQL conditions to use for each pharase, then determine which tables to use. A condition must have left hand side and right hand side, for example "A = B", "A in [1,2]". This is not a condition ```movie_popularity```, do not generate condition like this example. Please follow the format in the examples.
database schema :
table pages , columns = [ pages.words ( integer | values : 1081 , 68 ) , pages.page ( integer | values : 1 , 2 ) , pages.pid ( integer | primary key | comment : page id | values : 1 , 2 ) , pages.title ( text | values : Àbac , Abadia ) , pages.lid ( integer | comment : language id | values : 1 ) , pages.revision ( integer | values : 28236978 , 24086480 ) ]
table words , columns = [ words.word ( text | values : +,2 , +,33 ) , words.wid ( integer | primary key | comment : word id | values : 2148990 , 2506463 ) , words.occurrences ( integer | values : 242 , 16841 ) ]
table langs , columns = [ langs.pages ( integer | values : 1129144 ) , langs.words ( integer | values : 2764996 ) , langs.lid ( integer | primary key | comment : language id | values : 1 ) , langs.lang ( text | comment : language | values : ca ) , langs.locale ( text | values : ca_ES ) ]
table pages_words , columns = [ pages_words.pid ( integer | primary key | comment : page id | values : 1 , 2 ) , pages_words.wid ( integer | primary key | comment : word id | values : 1 , 2 ) , pages_words.occurrences ( integer | values : 30 , 8 ) ]
table langs_words , columns = [ langs_words.wid ( integer | primary key | comment : word id | values : 1 , 2 ) , langs_words.occurrences ( integer | values : 242 , 16841 ) , langs_words.lid ( integer | primary key | comment : language id | values : 1 ) ]
table biwords , columns = [ biwords.occurrences ( integer | values : 4 , 3 ) , biwords.lid ( integer | primary key | comment : language id | values : 1 ) , biwords.w1st ( integer | primary key | comment : word id of the first word | values : 1 , 2 ) , biwords.w2nd ( integer | primary key | comment : word id of the second word | values : 2 , 4 ) ]
foreign keys :
pages.lid = langs.lid
langs_words.wid = words.wid
langs_words.lid = langs.lid
pages_words.wid = words.wid
pages_words.pid = pages.pid
biwords.w2nd = words.wid
biwords.w1st = words.wid
biwords.lid = langs.lid

Matched contents are written in this format table.column (some values can be found in that column)
matched contents :
pages.words ( 1500 )
pages.page ( 1500 )
pages.pid ( 1500 )
pages.title ( Pages , 1500 )
words.word ( pages , words , calculates , differents , divides , percentages , counts , page's , wordes )
pages_words.occurrences ( 1500 )
langs_words.wid ( 1500 )
langs_words.occurrences ( 1500 )
biwords.occurrences ( 1500 )
biwords.w1st ( 1500 )
biwords.w2nd ( 1500 )

Question: Calculate the percentage of pages that have 1500 different words.
External knowledge: DIVIDE(COUNT(pages WHERE words = 1500), COUNT(pages)) as percentage;

Answer in this format:
Goal to select:
- The question asks for ['the percentage of pages'].
    + 'the percentage of pages': From the External knowledge `DIVIDE(COUNT(pages WHERE words = 1500), COUNT(pages)) as percentage`, so `the percentage of pages` refers to `COUNT(CASE WHEN pages.words = 1500 THEN 1 ELSE NULL END) * 100.0 / COUNT(pages.pid)`.
- The query should select: [`COUNT(CASE WHEN pages.words = 1500 THEN 1 ELSE NULL END) * 100.0 / COUNT(pages.pid)`]

Condition:
- 'pages that have 1500 different words': refers to column `pages.words`, so the condition is ```pages.words = 1500```.

Tables to use:
- From the goal to select and condition, the query should use tables [`pages`].

Final SQL query:
```
SELECT COUNT(CASE WHEN pages.words = 1500 THEN 1 ELSE NULL END) * 100.0 / COUNT(pages.pid) FROM pages;
```
=========
database schema :
table lists , columns = [ lists.list_followers ( integer | values : 5 , 1 ) , lists.list_update_timestamp_utc ( text | values : 2019-01-24 19:16:18 , 2018-12-03 15:12:20 ) , lists.list_url ( text ) , lists.list_id ( integer | primary key | values : 1 , 2 ) , lists.list_creation_timestamp_utc ( text | values : 2009-11-11 00:02:21 , 2009-11-11 00:05:11 ) , lists.list_title ( text | values : Headscratchers ) , lists.user_id ( integer | values : 88260493 , 45204418 ) , lists.list_description ( text ) , lists.list_movie_number ( integer | values : 5 , 3 ) , lists.list_comments ( integer | values : 3 , 2 ) ]
table lists_users , columns = [ lists_users.list_id ( integer | primary key | values : 192287 , 192313 ) , lists_users.list_update_date_utc ( text | values : 2019-11-26 , 2020-05-01 ) , lists_users.user_id ( integer | primary key | values : 2385 , 15264 ) , lists_users.list_creation_date_utc ( text | values : 2009-12-18 , 2010-01-30 ) , lists_users.user_subscriber ( integer | values : 1 , 0 ) , lists_users.user_trialist ( integer | values : 1 , 0 ) , lists_users.user_has_payment_method ( text | values : 1 , 0 ) , lists_users.user_eligible_for_trial ( text | values : 0 , 1 ) , lists_users.user_avatar_image_url ( text ) , lists_users.user_cover_image_url ( text ) ]
table movies , columns = [ movies.director_name ( text | values : Esteban Sapir , Oskar Roehler ) , movies.movie_title ( text | values : La Antena , Elementary Particles ) , movies.director_id ( text | values : 131 , 73 ) , movies.movie_url ( text ) , movies.movie_id ( integer | primary key | values : 1 , 2 ) , movies.director_url ( text ) , movies.movie_popularity ( integer | values : 105 , 23 ) , movies.movie_image_url ( text ) , movies.movie_release_year ( integer | values : 2007 , 2006 ) , movies.movie_title_language ( text | values : en ) ]
table ratings_users , columns = [ ratings_users.user_id ( integer | values : 41579158 , 68654088 ) , ratings_users.rating_date_utc ( text | values : 2017-06-10 , 2012-10-02 ) , ratings_users.user_subscriber ( integer | values : 0 , 1 ) , ratings_users.user_trialist ( integer | values : 0 , 1 ) , ratings_users.user_eligible_for_trial ( integer | values : 1 , 0 ) , ratings_users.user_has_payment_method ( integer | values : 0 , 1 ) , ratings_users.user_cover_image_url ( text ) , ratings_users.user_avatar_image_url ( text ) ]
table ratings , columns = [ ratings.critic ( text ) , ratings.rating_timestamp_utc ( text | values : 2017-06-10 12:38:33 , 2014-08-15 23:42:31 ) , ratings.user_id ( integer | values : 41579158 , 85981819 ) , ratings.rating_id ( integer | values : 15610495 , 10704606 ) , ratings.rating_score ( integer | values : 3 , 2 ) , ratings.critic_comments ( integer | values : 0 , 2 ) , ratings.movie_id ( integer | values : 1066 , 1067 ) , ratings.critic_likes ( integer | values : 0 , 1 ) , ratings.rating_url ( text ) , ratings.user_subscriber ( integer | values : 0 , 1 ) ]
foreign keys :
lists.user_id = lists_users.user_id
ratings_users.user_id = lists_users.user_id
lists_users.user_id = lists.user_id
lists_users.list_id = lists.list_id
ratings.user_id = ratings_users.user_id
ratings.rating_id = ratings.rating_id
ratings.user_id = lists_users.user_id
ratings.movie_id = movies.movie_id

Matched contents are written in this format table.column (some values can be found in that column)
matched contents :
lists.list_id ( 2012 )
lists.list_title ( on MUBI , on Mubi , Mubi , MUBI , The List. )

Question: What are the URL to the list page on Mubi of the lists with followers between 1-2 and whose last update timestamp was on 2012?
External knowledge: URL to the list page on Mubi refers to list_url; list_followers = 1 OR list_followers = 2; last update timestamp was on 2012 refers to list_update_timestamp_utc BETWEEN '2012-1-1' AND '2012-12-31'; 

Answer in this format:
Goal to select:
- The question asks for ['the URL to the list page on Mubi'].
    + 'the URL to the list page on Mubi', from the External knowledge, 'URL to the list page on Mubi refers to list_url', so 'the URL to the list page on Mubi' refers to `lists.list_url`.

- The query should select: [`lists.list_url`]

Condition:
- 'with followers between 1-2': check with the external knowledge `list_followers = 1 OR list_followers = 2`, so `with followers between 1-2` refers to column `lists.list_followers`, the condition is `lists.list_followers BETWEEN 1 AND 2`.
- 'last update timestamp was on 2012': check with the external knowledge `list_update_timestamp_utc BETWEEN '2012-1-1' AND '2012-12-31'` refers to column `lists.list_update_timestamp_utc`. The condition is `lists.list_update_timestamp_utc BETWEEN '2012-01-01' AND '2012-12-31'`.

Tables to use:
- From the goal to select and condition, the query should use tables [`lists`].

Final SQL query:
```
SELECT lists.list_url 
FROM lists 
WHERE lists.list_followers BETWEEN 1 AND 2 
AND lists.list_update_timestamp_utc BETWEEN '2012-01-01' AND '2012-12-31';
```
=========
database schema :
table matchs , columns = [ matchs.div ( text | comment : division | values : B1 , D1 ) , matchs.ftr ( text | comment : final-time results | values : A , D ) , matchs.hometeam ( text | values : Club Brugge , Antwerp ) , matchs.season ( integer | values : 2021 , 2020 ) , matchs.awayteam ( text | values : Charleroi , Mouscron ) , matchs.date ( date | values : 2020-08-08 , 2020-08-09 ) , matchs.fthg ( integer | comment : final-time home-team goals | values : 0 , 1 ) , matchs.ftag ( integer | comment : final-time away-team goals | values : 1 , 0 ) ]
table divisions , columns = [ divisions.division ( text | primary key | values : B1 , D1 ) , divisions.country ( text | values : Belgium , Deutschland ) , divisions.name ( text | values : Division 1A , Bundesliga ) ]
foreign keys :
matchs.div = divisions.division

Matched contents are written in this format table.column (some values can be found in that column)
matched contents :
matchs.season ( 2017 )
divisions.country ( Spain )
divisions.name ( LaLiga )

Question: From the Spanish LaLiga division in the 2017 season, which team won the most times as a local team and by what percentage?
External knowledge: local team refers to hometeam; Spanish means belong to the country = 'Spain'; LaLiga is a name of division; won as a local team refers to ftr = 'H', where H stands for home victory; divIDE(COUNT(div where name = 'LaLiga', country = 'Spain', season = 2017, FRT = 'H'), COUNT(div where name = 'LaLiga', country = 'Spain', season = 2017)) as percentage;

Answer in this format:
Goal to select:
- The question asks for ['which team won the most times as a local team', 'by what percentage'].
    + 'which team won the most times as a local team': From the external knowledge, `local team refers to hometeam`, so this needs to select 'hometeam'.
    + 'by what percentage': From the external knowledge `divIDE(COUNT(div where name = 'LaLiga', country = 'Spain', season = 2017, FRT = 'H'), COUNT(div where name = 'LaLiga', country = 'Spain', season = 2017))`, so this needs to select `COUNT(CASE WHEN divisions.name = 'LaLiga' AND divisions.country = 'Spain' AND matchs.season = 2017 AND matchs.ftr = 'H' THEN 1 ELSE NULL END) * 100.0 / COUNT(CASE WHEN divisions.name = 'LaLiga' AND divisions.country = 'Spain' AND matchs.season = 2017 THEN 1 ELSE NULL END)`.
- The query should select: [`hometeam`, `COUNT(CASE WHEN divisions.name = 'LaLiga' AND divisions.country = 'Spain' AND matchs.season = 2017 AND matchs.ftr = 'H' THEN 1 ELSE NULL END) * 100.0 / COUNT(CASE WHEN divisions.name = 'LaLiga' AND divisions.country = 'Spain' AND matchs.season = 2017 THEN 1 ELSE NULL END)`].
- 'won the most times': The query also needs to order the result by percentage and get top-1 result.

Condition:
- 'From the Spanish': From external knowledge, `Spanish means belong to the country = 'Spain'`, this refers to the column `divisions.country`, so the condition is `divisions.country = 'Spain'`.
- 'LaLiga division': From external knowledge `LaLiga is a name of division`, this refers to column 'divisions.name', so the condition is `divisions.name = 'LaLiga'`.
- '2017 season': From external knowledge `season = 2017`, this refers to column 'matchs.season', so the condition `matchs.season = 2017`.
- 'won as a local team': From external knowledge `won as a local team refers to ftr = 'H'`, so this refers to column 'matchs.ftr', so the condition is `matchs.ftr = 'H'`.

Tables to use:
- From the goal to select and condition, the query should use tables [`matchs`, `divisions`].

Final SQL query:
```
SELECT matchs.hometeam, 
       COUNT(CASE WHEN divisions.name = 'LaLiga' AND divisions.country = 'Spain' AND matchs.season = 2017 AND matchs.ftr = 'H' THEN 1 ELSE NULL END) * 100.0 / COUNT(CASE WHEN divisions.name = 'LaLiga' AND divisions.country = 'Spain' AND matchs.season = 2017 THEN 1 ELSE NULL END) AS percentage 
FROM matchs 
JOIN divisions ON matchs.div = divisions.division 
WHERE divisions.country = 'Spain' 
  AND divisions.name = 'LaLiga' 
  AND matchs.season = 2017
ORDER BY percentage DESC LIMIT 1;
```
=========
database schema :
table movies , columns = [ movies.movie_release_year ( integer | values : 2007 , 2006 ) , movies.movie_title ( text | values : La Antena , Elementary Particles ) , movies.movie_popularity ( integer | values : 105 , 23 ) , movies.movie_id ( integer | primary key | values : 1 , 2 ) , movies.movie_title_language ( text | values : en ) , movies.director_name ( text | values : Esteban Sapir , Oskar Roehler ) , movies.movie_url ( text ) , movies.movie_image_url ( text ) , movies.director_id ( text | values : 131 , 73 ) , movies.director_url ( text ) ]
table ratings_users , columns = [ ratings_users.user_id ( integer | values : 41579158 , 68654088 ) , ratings_users.user_subscriber ( integer | values : 0 , 1 ) , ratings_users.user_trialist ( integer | values : 0 , 1 ) , ratings_users.user_has_payment_method ( integer | values : 0 , 1 ) , ratings_users.rating_date_utc ( text | values : 2017-06-10 , 2012-10-02 ) , ratings_users.user_cover_image_url ( text ) , ratings_users.user_eligible_for_trial ( integer | values : 1 , 0 ) , ratings_users.user_avatar_image_url ( text ) ]
table lists_users , columns = [ lists_users.list_id ( integer | primary key | values : 192287 , 192313 ) , lists_users.user_id ( integer | primary key | values : 2385 , 15264 ) , lists_users.user_trialist ( integer | values : 1 , 0 ) , lists_users.user_has_payment_method ( text | values : 1 , 0 ) , lists_users.user_subscriber ( integer | values : 1 , 0 ) , lists_users.user_eligible_for_trial ( text | values : 0 , 1 ) , lists_users.user_cover_image_url ( text ) , lists_users.user_avatar_image_url ( text ) , lists_users.list_creation_date_utc ( text | values : 2009-12-18 , 2010-01-30 ) , lists_users.list_update_date_utc ( text | values : 2019-11-26 , 2020-05-01 ) ]
table lists , columns = [ lists.list_title ( text | values : Headscratchers ) , lists.list_movie_number ( integer | values : 5 , 3 ) , lists.list_id ( integer | primary key | values : 1 , 2 ) , lists.user_id ( integer | values : 88260493 , 45204418 ) , lists.list_description ( text ) , lists.list_comments ( integer | values : 3 , 2 ) , lists.list_url ( text ) , lists.list_followers ( integer | values : 5 , 1 ) , lists.list_third_image_url ( text ) , lists.list_second_image_url ( text ) ]
table ratings , columns = [ ratings.movie_id ( integer | values : 1066 , 1067 ) , ratings.rating_id ( integer | values : 15610495 , 10704606 ) , ratings.critic ( text ) , ratings.user_id ( integer | values : 41579158 , 85981819 ) , ratings.rating_score ( integer | values : 3 , 2 ) , ratings.critic_comments ( integer | values : 0 , 2 ) , ratings.critic_likes ( integer | values : 0 , 1 ) , ratings.rating_url ( text ) , ratings.user_trialist ( integer | values : 0 , 1 ) , ratings.user_subscriber ( integer | values : 0 , 1 ) ]
foreign keys :
lists.user_id = lists_users.user_id
ratings_users.user_id = lists_users.user_id
lists_users.user_id = lists.user_id
lists_users.list_id = lists.list_id
ratings.user_id = ratings_users.user_id
ratings.rating_id = ratings.rating_id
ratings.user_id = lists_users.user_id
ratings.movie_id = movies.movie_id

Matched contents are written in this format table.column (some values can be found in that column)
matched contents :
movies.movie_release_year ( 1945 )
movies.movie_title ( Year , 1945 , Order , The Years , Release )
movies.movie_id ( 1945 )
lists_users.list_id ( 1945 )
lists.list_title ( 1945 , Sort , Titles. , title , Title )
lists.list_id ( 1945 )
ratings.movie_id ( 1945 )
ratings.rating_id ( 1945 )

Question: Sort the listing by the descending order of movie popularity.
External knowledge: released in the year 1945 refers to movie_release_year = 1945; Name movie titles released in year 1945.

Answer in this format:
Goal to select:
- The question asks for ['movie titles', 'sorted by popularity in descending order'].
    + 'movie titles': this refers to `movies.movie_title`.

- The query should select: [`movies.movie_title`]
- The query needs to `sort by the descending order of movie popularity`, this refers to `ORDER BY movies.movie_popularity DESC`.

Condition:
- 'released in the year 1945': from external knowledge `released in the year 1945`, so this refers to column `movies.movie_release_year`, so the condition is `movies.movie_release_year = 1945`.

Tables to use:
- From the goal to select and condition, the query should use tables [`movies`].

Final SQL query:
```
SELECT movies.movie_title 
FROM movies 
WHERE movies.movie_release_year = 1945 
ORDER BY movies.movie_popularity DESC;
```