批量插入期间的SQL Server表锁定

Question

Below is the sample query, consider A

INSERT INTO Target (Col1,Col2,Col3,Col4) ----------------Statement#1
Select A.Col1,B.Col2,A.Col3,C.Col4       ----------------Statement#2
FROM A WITH(NOLOCK) INNER JOIN B WITH(NOLOCK)
    ON A.Id = B.ID
    LEFT JOIN C WITH NOLOCK
    ON C.Id = B.ID
Where A.Id = 11

At which stage the lock will be applied on table [exclusive lock?], how SQL is going to execute the query?

1. Result will be fetched from table A, B and C based on join and where clause.
2. On ready result, start inserting data in table and at same time apply the lock on table.

So when actual data is written on the page table is locked but not during select even though it is INSERT INTO with SELECT?

Answer 1

Those two steps are the logical steps for query execution. What SQL Server can do/do at physical level is another story. At this moment:

INSERT INTO Target (Col1,Col2,Col3,Col4) ----------------Statement#1
Select A.Col1,B.Col2,A.Col3,C.Col4       ----------------Statement#2
FROM A WITH(NOLOCK) INNER JOIN B WITH(NOLOCK)
    ON A.Id = B.ID
    LEFT JOIN C WITH NOLOCK
    ON C.Id = B.ID
Where A.Id = 11

for every output record (see SELECT clause) it takes an X lock on a RID or a KEY within target table ( RID for heap / KEY for clustered index) and it inserts that record. This steps are repeated for every output record. So, it doesn't read all records from source tables and only after this step it starts inserting records into target table. Because of NOLOCK table hint on source table it will takes only Sch-S (schema stability) locks on these tables.

If you want to take an X lock on target table then you could use

INSERT INTO Target WITH(TABLOCKX) (Col1,Col2,Col3,Col4)
SELECT ...

If you want minimally logged inserts then please read this article .

Answer 2

Did you specify any "Table Lock" hint. If you want to Row-level lock Set "Table Lock" to off.

or check this it will help you...

http://technet.microsoft.com/en-us/library/ms180876(v=sql.105).aspx