比较两个“兄弟”SQL表的结构

Question

为了比较基表和它的存档表（通常是arc_ + 基表名称），我编写了以下内容来显示两个表中的列数，即存在于一侧但不存在的列另一方面，以及两边不同的列：

DECLARE @base_tbl varchar(20) = 'appointments'

DECLARE @arch_tbl varchar(20) = 'arc_' + @base_tbl
DECLARE @schema varchar(35) = 'some'

-- Count in base table.
SELECT @base_tbl, COUNT(*)
FROM information_schema.columns
WHERE TABLE_NAME = @base_tbl AND TABLE_SCHEMA = @schema

-- Count in archive table.
SELECT @arch_tbl, COUNT(*)
FROM information_schema.columns
WHERE TABLE_NAME = @arch_tbl AND TABLE_SCHEMA = @schema

-- Columns only in the base or in the archive table.
SELECT TABLE_NAME, COLUMN_NAME, DATA_TYPE, CHARACTER_MAXIMUM_LENGTH, NUMERIC_PRECISION, IS_NULLABLE
FROM information_schema.columns
WHERE TABLE_NAME IN (@base_tbl, @arch_tbl) AND TABLE_SCHEMA = @schema
  AND COLUMN_NAME IN (
      SELECT DISTINCT COLUMN_NAME FROM
      (
          SELECT TABLE_NAME, COLUMN_NAME
          FROM information_schema.columns
          WHERE TABLE_NAME = @base_tbl AND TABLE_SCHEMA = @schema
              UNION
          SELECT TABLE_NAME, COLUMN_NAME
          FROM information_schema.columns
          WHERE TABLE_NAME = @arch_tbl AND TABLE_SCHEMA = @schema
      ) u
      GROUP BY COLUMN_NAME
      HAVING COUNT(*) = 1)
ORDER BY COLUMN_NAME, TABLE_NAME

-- Differences.
SELECT TABLE_NAME, COLUMN_NAME, DATA_TYPE, CHARACTER_MAXIMUM_LENGTH, NUMERIC_PRECISION, IS_NULLABLE
FROM information_schema.columns
WHERE TABLE_NAME IN (@base_tbl, @arch_tbl) AND TABLE_SCHEMA = @schema
  AND COLUMN_NAME IN (
      SELECT DISTINCT COLUMN_NAME FROM
      (
          SELECT COLUMN_NAME, DATA_TYPE, CHARACTER_MAXIMUM_LENGTH, NUMERIC_PRECISION
          FROM information_schema.columns
          WHERE TABLE_NAME = @base_tbl AND TABLE_SCHEMA = @schema
              UNION
          SELECT COLUMN_NAME, DATA_TYPE, CHARACTER_MAXIMUM_LENGTH, NUMERIC_PRECISION
          FROM information_schema.columns
          WHERE TABLE_NAME = @arch_tbl AND TABLE_SCHEMA = @schema
      ) v
GROUP BY COLUMN_NAME
HAVING COUNT(*) > 1)
ORDER BY COLUMN_NAME, TABLE_NAME

我想知道这是否可以改进，以提高可读性或更好的输出支持。 WDYT？

Answer 1

FULL JOIN 不做这一切吗？

SELECT c1.COLUMN_NAME, c1.DATA_TYPE, c2.COLUMN_NAME, c2.DATA_TYPE
FROM 
(
    SELECT *
    FROM information_schema.columns c1
    WHERE c1.TABLE_NAME = 'xxx' AND c1.table_schema = 'xxx'
) c1
FULL JOIN 
(
    SELECT *
    FROM information_schema.columns c2
    WHERE c2.TABLE_NAME = 'yyy' AND c2.table_schema = 'yyy'
) c2
ON c1.COLUMN_NAME = c2.COLUMN_NAME
ORDER BY
    CASE WHEN c1.COLUMN_NAME IS NULL OR c2.COLUMN_NAME IS NULL THEN 0 ELSE 1 END,
    ISNULL(c1.COLUMN_NAME, c2.COLUMN_NAME)