需要帮助旋转数据

Posted 2023-05-09

【中文标题】需要帮助旋转数据

【英文标题】：Need assistance pivoting data

【发布时间】：2020-10-29 20:53:05

【问题描述】：

我查看了所有示例，但无法完全理解 Oracle 的 PIVOT 功能。我也不确定是否可以在单个查询中完成我需要的操作。

我正在尝试处理一些选举数据。以下是一些示例数据：

create table elections_test (election_date date, county_name varchar2(40), precinct_code varchar2(7), race_number varchar2(7), race_description varchar2(250), cand_name varchar2(100), cand_party varchar2(3), cand_tot_votes number, cand_absmail_votes number, cand_early_votes number, cand_elecday_votes number);

insert into elections_test values ('8-NOV-2016', 'CREEK', '190420', '10001', 'FOR PRESIDENT AND VICE PRESIDENT', 'DONALD TRUMP', 'REP', 402, 16, 42, 344);
insert into elections_test values ('8-NOV-2016', 'CREEK', '190420', '10001', 'FOR PRESIDENT AND VICE PRESIDENT', 'GARY JOHNSON', 'LIB', 33, 1, 2, 30);
insert into elections_test values ('8-NOV-2016', 'CREEK', '190420', '10001', 'FOR PRESIDENT AND VICE PRESIDENT', 'HILLARY CLINTON', 'DEM', 155, 15, 35, 105);
insert into elections_test values ('8-NOV-2016', 'CREEK', '190420', '10003', 'FOR US SENATOR', 'JAMES BOND', 'REP', 438, 16, 43, 379);
insert into elections_test values ('8-NOV-2016', 'CREEK', '190420', '10003', 'FOR US SENATOR', 'MIKE SMITH', 'DEM', 126, 11, 33, 82);
insert into elections_test values ('8-NOV-2016', 'CREEK', '190420', '10003', 'FOR US SENATOR', 'ROBERT JONES', 'IND', 7, 1, 0, 6);
insert into elections_test values ('8-NOV-2016', 'CREEK', '190420', '10003', 'FOR US SENATOR', 'JOHN JAMES', 'LIB', 17, 1, 1, 15);
insert into elections_test values ('8-NOV-2016', 'ALFALFA', '020121', '10001', 'FOR PRESIDENT AND VICE PRESIDENT', 'DONALD TRUMP', 'REP', 300, 20, 40, 240);
insert into elections_test values ('8-NOV-2016', 'ALFALFA', '020121', '10001', 'FOR PRESIDENT AND VICE PRESIDENT', 'GARY JOHNSON', 'LIB', 23, 1, 1, 21);
insert into elections_test values ('8-NOV-2016', 'ALFALFA', '020121', '10001', 'FOR PRESIDENT AND VICE PRESIDENT', 'HILLARY CLINTON', 'DEM', 125, 10, 20, 95);
insert into elections_test values ('8-NOV-2016', 'ALFALFA', '020121', '10003', 'FOR US SENATOR', 'JAMES BOND', 'REP', 330, 10, 3, 317);
insert into elections_test values ('8-NOV-2016', 'ALFALFA', '020121', '10003', 'FOR US SENATOR', 'MIKE SMITH', 'DEM', 100, 10, 8, 82);
insert into elections_test values ('8-NOV-2016', 'ALFALFA', '020121', '10003', 'FOR US SENATOR', 'ROBERT JONES', 'IND', 4, 1, 0, 3);
insert into elections_test values ('8-NOV-2016', 'ALFALFA', '020121', '10003', 'FOR US SENATOR', 'JOHN JAMES', 'LIB', 12, 0, 1, 11);

这是我想要的：

election_date | county_name | precinct_code | race_number | race_description                 | total_votes | total_absmail_votes | total_early_votes | total_elecday_votes | rep_total_votes | rep_absmail_votes | rep_early_votes | rep_elecday_votes | dem_total_votes | dem_absmail_votes | dem_early_votes | dem_elecday_votes | other_total_votes | other_absmail_votes | other_early_votes | other_elecday_votes
8-NOV-2016    | CREEK       | 190420        | 10001       | FOR PRESIDENT AND VICE PRESIDENT | 590         | 32                  | 79                | 479                 | 402             | 16                | 42              | 344               | 155             | 15                | 35              | 105               | 33                | 1                   | 2                 | 30
8-NOV-2016    | CREEK       | 190420        | 10003       | FOR US SENATE                    | ###         | ##                  | #                 | #                   | ###             | ##                | ##              | ###               | #               | #                 | ##              | ###               | ##                | #                   | #                 | ##
8-NOV-2016    | ALFALFA     | 020121        | 10001       | FOR PRESIDENT AND VICE PRESIDENT | ###         | ##                  | #                 | #                   | ###             | ##                | ##              | ###               | #               | #                 | ##              | ###               | ##                | #                   | #                 | ##
8-NOV-2016    | ALFALFA     | 020121        | 10003       | FOR US SENATE                    | ###         | ##                  | #                 | #                   | ###             | ##                | ##              | ###               | #               | #                 | ##              | ###               | ##                | #                   | #                 | ##

基本上只是试图将每个选举日期、选区代码和种族都压平为一个记录。任何帮助将不胜感激。

【参考方案1】：

您可以使用条件聚合：

SELECT e.election_date,
       e.county_name,
       e.precinct_code,
       e.race_number,
       e.race_description,
       SUM(e.cand_tot_votes) AS total_votes,
       SUM(e.cand_absmail_votes) AS total_absmail_votes,
       SUM(e.cand_early_votes) AS total_early_votes,
       SUM(e.cand_elecday_votes) AS total_elecday_votes,
       SUM(CASE WHEN e.cand_party = 'REP' THEN e.cand_tot_votes END) AS rep_total_votes,
       SUM(CASE WHEN e.cand_party = 'REP' THEN e.cand_absmail_votes END) AS rep_absmail_votes,
       SUM(CASE WHEN e.cand_party = 'REP' THEN e.cand_early_votes END) AS rep_early_votes,
       SUM(CASE WHEN e.cand_party = 'REP' THEN e.cand_elecday_votes END) AS rep_elecday_votes,
       SUM(CASE WHEN e.cand_party = 'DEM' THEN e.cand_tot_votes END) AS dem_total_votes,
       SUM(CASE WHEN e.cand_party = 'DEM' THEN e.cand_absmail_votes END) AS dem_absmail_votes,
       SUM(CASE WHEN e.cand_party = 'DEM' THEN e.cand_early_votes END) AS dem_early_votes,
       SUM(CASE WHEN e.cand_party = 'DEM' THEN e.cand_elecday_votes END) AS dem_elecday_votes       
  FROM elections_test e
 GROUP BY e.election_date,
          e.county_name,
          e.precinct_code,
          e.race_number,
          e.race_description

为了根据需要旋转您的数据。

Demo

【讨论】：

您是说从该查询开始，然后将其用作 PIVOT 的基础吗？不要以为那行得通。我没有政党信息，因此无法按政党细分选票。

没有按党派细分。

对不起，我最近能理解你的意思@sukach，只是将这些案例添加到答案中。