或者换一种说法:如何找到表中的哪条记录具有最少的空字段或空白字段?
无需单独计算每个字段 - 该表有 161 个字段。
这可以通过构造一个动态查询来完成,例如:
SELECT id,
(IF(col1 = NULL OR col1 = "", 1, 0) +
IF(col2 = NULL OR col3 = "", 1, 0) +
...
IF(coln = NULL OR coln = "", 1, 0)
) AS null_count
FROM table_name
ORDER BY null_count DESC
LIMIT 1;
这可以通过使用INFORMATION_SCHEMA.COLUMNS
然后执行形成一个新的动态查询来轻松完成dynamic sql query
。此外,您可能需要通过将会话GROUP_CONCAT
级别变量设置group_concat_max_len
为更高的值来增加函数输出的最大长度。
SET GLOBAL group_concat_max_len = 4294967295;
SELECT @query1 := CONCAT('SELECT id,
(',
GROUP_CONCAT(CONCAT('IF(',COLUMN_NAME,' IS NULL OR ',
COLUMN_NAME,' = "", 1, 0
)
')
SEPARATOR ' + '),
') AS null_count
FROM table_name
ORDER BY null_count DESC
LIMIT 1')
FROM INFORMATION_SCHEMA.COLUMNS
WHERE TABLE_SCHEMA = SCHEMA()
AND TABLE_NAME = 'table_name';
PREPARE stmt FROM @query1; EXECUTE stmt; DEALLOCATE PREPARE stmt;
免责声明:我的问题错了,以为 OP 正在寻找具有最多 NULL 的表的列。尽管如此,也许它对任何人都有用。
像这样创建一个过程:
drop procedure if exists test_most_pop_field;
DELIMITER $$
CREATE PROCEDURE test_most_pop_field(IN tableName varchar(100))
BEGIN
DECLARE done INT DEFAULT 0;
DECLARE sql_query VARCHAR(255);
DECLARE cur CURSOR FOR
SELECT CONCAT('INSERT INTO tmp_result(columnName, numberOfEmptyRows) SELECT "', COLUMN_NAME, '" AS columnName, SUM(IF(',COLUMN_NAME,' IS NULL OR ', COLUMN_NAME,' = "", 1, 0)) AS numberEmptyRows FROM ', TABLE_NAME)
FROM INFORMATION_SCHEMA.COLUMNS
WHERE TABLE_SCHEMA = SCHEMA()
AND TABLE_NAME = tableName;
DECLARE CONTINUE HANDLER FOR SQLSTATE '02000' SET done = 1;
DROP TABLE IF EXISTS tmp_result;
CREATE TEMPORARY TABLE tmp_result(columnName varchar(100), numberOfEmptyRows int);
OPEN cur;
REPEAT
FETCH cur INTO sql_query;
IF NOT done THEN
BEGIN
SET @sql = sql_query; /*this extra step is necessary, cause otherwise it's a syntax error, don't ask me why*/
PREPARE stmt FROM @sql;
EXECUTE stmt;
DEALLOCATE PREPARE stmt;
END;
END IF;
UNTIL done END REPEAT;
CLOSE cur;
SELECT * FROM tmp_result ORDER BY numberOfEmptyRows DESC /*optionally LIMIT 1*/;
END $$
DELIMITER ;
然后使用您要检查的表名调用它:
CALL test_most_pop_field('yourTableName');