开发者

How can I find all columns that are filled with 100% nulls in my SQL Server database schema?

Is there a SQL way to find out which columns in my schema are completely full of null values? There are some fields in a couple of tables that I know are not used by the application and will be开发者_如何学运维 dropped, but I wanted to see if there was an automated way/script to find this out across the whole database to find candidates for code review/possible removal.

Running SQL Server 2005 on x86 if it matters.

Thanks in advance!


create table #SuspectColumns (
    TABLE_SCHEMA sysname,
    TABLE_NAME sysname,
    COLUMN_NAME sysname
)

declare csrColumns cursor fast_forward for
    select TABLE_SCHEMA, TABLE_NAME, COLUMN_NAME
        from INFORMATION_SCHEMA.COLUMNS
        where IS_NULLABLE = 'YES'

declare @TABLE_SCHEMA sysname,
        @TABLE_NAME sysname,
        @COLUMN_NAME sysname,
        @sql nvarchar(max)  

open csrColumns

while (1=1) begin
    fetch next
        from csrColumns
        into @TABLE_SCHEMA, @TABLE_NAME, @COLUMN_NAME

    if @@FETCH_STATUS<>0 break

    set @sql = N'if not exists(select 1 from ' + QUOTENAME(@TABLE_SCHEMA) + N'.' + QUOTENAME(@TABLE_NAME) + N' where ' + QUOTENAME(@COLUMN_NAME) + N'is not null)
                     insert into #SuspectColumns values (''' + @TABLE_SCHEMA + N''',''' + @TABLE_NAME + N''',''' + @COLUMN_NAME + N''')'

    exec sp_executesql @sql
end /* while */

close csrColumns
deallocate csrColumns

select * from #SuspectColumns

drop table #SuspectColumns


you can return the max(column) and check for nulls


Off the top of my head, I believe the following SQL should work. It will execute a query for each table/column combination and the query will return the table name and column name if that table/column combination either has no rows or all null rows.

DECLARE @table_columns TABLE
(
  table_name nvarchar(128),
  column_name nvarchar(128)
);
DECLARE @table_name nvarchar(128);
DECLARE @column_name nvarchar(128);

INSERT INTO @table_columns(table_name, column_name)
select TABLE_NAME, COLUMN_NAME from INFORMATION_SCHEMA.columns;

while (select count(*) from @table_columns) > 0
begin
    select top 1 @table_name = table_name, @column_name = column_name from @table_columns
    exec('SELECT ''' + @table_name + ''' as table_name, ''' + @column_name + ''' as column_name WHERE NOT EXISTS (SELECT TOP 1 * FROM ' + @table_name + ' WHERE ' + @column_name + ' IS NOT NULL)')
    delete from @table_columns where table_name = @table_name and column_name = @column_name 
end
0

上一篇:

下一篇:

精彩评论

暂无评论...
验证码 换一张
取 消

最新问答

问答排行榜