SQL Server 插入包含“时间戳（行版本）”列的表

Question

I'm creating a C# Winforms application for recipe management in an industrial environment.

I created a SQL Server table with 130 columns. The table contains a column called CheckData (of datatype Timestamp ), which I use to detect changes made to a row.

If I insert a new row to that table all works fine. The code I use is:

INSERT INTO tablename (Column1, column2, column3, column4) 
VALUES (value1, value2, value3, value4)

I just assign values to major columns, the others get their default value. I do not assign a value to the timestamp field since it's written by the system.

Additionally, I want to copy a row from this table to the same table (duplicate a data record).

I copy the source row to a temporary table, drop the ID (primary key) and the timestamp fields in that temporary table and try to insert that only row in the temporary table into the table. This fails.

Here's the code:

SELECT * 
INTO #temptable 
FROM tablename 
WHERE Recipe_No = 8;

ALTER TABLE #temptable DROP COLUMN ID, CHECKDATA;

ALTER TABLE #temptable REBUILD;

UPDATE #temptable 
SET Recipe_No = 9, Recipe_Name = 'Test' 
WHERE Recipe_No = 8;

INSERT INTO tablename 
    SELECT * FROM #temptable;

I don't understand where the difference is between inserting a new row thru INSERT INTO xxx (yyy) VALUES (zzz) and INSERT INTO xxx SELECT * FROM yyy . In both cases I don't try to write the timestamp value in the new row.

Does anybody have an idea what I'm missing here?

Answer 1

I don't understand where the difference is between inserting a new row thru INSERT INTO xxx (yyy) VALUES (zzz) and INSERT INTO xxx SELECT * FROM yyy.

With this,

INSERT INTO xxx SELECT * FROM yyy.

you are failing to specify the column mappings from the SELECT to the target table. You should always use

INSERT INTO xxx (Column1, Column2, . . .)
SELECT (Column1, Column2, . . .)
FROM yyy

Here's a simplified example of what you're attempting:

drop table if exists t

create table t(id int, a int)
insert into t(id,a) values (1,1)

select * into #t from t where id = 1

alter table #t drop column id

insert into t select * from #t 

and it will fail with

Msg 213, Level 16, State 1, Line 12
Column name or number of supplied values does not match table definition.

because the temp table doesn't even have the same number of columns. And even if it did, you wouldn't know for sure that the column mappings were correct.

Answer 2

It is failing because essentially your command

INSERT INTO tablename SELECT * FROM #temptable;";

Is telling SQL - "Insert everything into this table from this temp table."

While you can work around this, I would say why don't you just try inserting into only the columns made available in your current table with only the values you would like to include. Instead of needing to drop the columns/values, you just don't import it to begin with.

An alternative - if you can write to a helper table, it may be beneficial to INSERT INTO that helper table, as opposed to a temp table, the values you have. Then transform that helper table, and THEN you can do INSERT INTO final_table SELECT * FROM helper . This should give you the results you're looking for.

I hope this is helpful, and I hope it explains why your current command is failing.