“选择1”是做什么的?

我读了一些文章,但真的不明白 select 1 from是做什么的?有人说“你应该使用 select 1而不是 select *”。下面是一个示例表:

cust_id     cust_name       cust_address


1000000001  Village Toys    Mapl
1000000002  Kids Place      South
1000000003  Fun4All         Sunny
1000000004  Fun4All         Riverside
1000000005  The Toy Store   53rd

当我写 select 1 from customer_table的时候,结果是什么? 这个语句做了什么?

194789 次浏览

The construction is usually used in "existence" checks

if exists(select 1 from customer_table where customer = 'xxx')

or

if exists(select * from customer_table where customer = 'xxx')

Both constructions are equivalent. In the past people said the select * was better because the query governor would then use the best indexed column. This has been proven not true.

It does what you ask, SELECT 1 FROM table will SELECT (return) a 1 for every row in that table, if there were 3 rows in the table you would get

1
1
1

Take a look at Count(*) vs Count(1) which may be the issue you were described.

The statement SELECT 1 FROM SomeTable just returns a column containing the value 1 for each row in your table. If you add another column in, e.g. SELECT 1, cust_name FROM SomeTable then it makes it a little clearer:

            cust_name
----------- ---------------
1           Village Toys
1           Kids Place
1           Fun4All
1           Fun4All
1           The Toy Store
select 1 from table

will return a column of 1's for every row in the table. You could use it with a where statement to check whether you have an entry for a given key, as in:

if exists(select 1 from table where some_column = 'some_value')

What your friend was probably saying is instead of making bulk selects with select * from table, you should specify the columns that you need precisely, for two reasons:

1) performance & you might retrieve more data than you actually need.

2) the query's user may rely on the order of columns. If your table gets updated, the client will receive columns in a different order than expected.

SELECT COUNT(*) in EXISTS/NOT EXISTS

EXISTS(SELECT CCOUNT(*) FROM TABLE_NAME WHERE CONDITIONS) - the EXISTS condition will always return true irrespective of CONDITIONS are met or not.

NOT EXISTS(SELECT CCOUNT(*) FROM TABLE_NAME WHERE CONDITIONS) - the NOT EXISTS condition will always return false irrespective of CONDITIONS are met or not.

SELECT COUNT 1 in EXISTS/NOT EXISTS

EXISTS(SELECT CCOUNT 1 FROM TABLE_NAME WHERE CONDITIONS) - the EXISTS condition will return true if CONDITIONS are met. Else false.

NOT EXISTS(SELECT CCOUNT 1 FROM TABLE_NAME WHERE CONDITIONS) - the NOT EXISTS condition will return false if CONDITIONS are met. Else true.