How do I rewrite this MySQL query so it doesn't throw this error: You can't specify target table 'crawlLog' for update in FROM clause?

Question

I'm trying to get an id from a companies table where the id is not yet in the crawlLog table. Then I need to insert that companyId into the crawlLog table.

I need to do this in one call so that parallel crawlers don't pull the same url after some other crawler has selected a url, but hasn't inserted it into the crawl log yet. I don't want to lock tables because of other problems that generates.

I get this error from both queries below:

You can't specify target table 'crawlLog' for update in FROM clause

Here are two queries i've tried to do the same thing.

INSERT INTO crawlLog (companyId, timeStartCrawling)
VALUES
(
    (
        SELECT companies.id FROM companies
        LEFT OUTER JOIN crawlLog
        ON companies.id = crawlLog.companyId
        WHERE crawlLog.companyId IS NULL
        LIMIT 1
    ),
    now()
)

I've also tried this, but get the same error:

INSERT INTO crawlLog (companyId, timeStartCrawling)
VALUES
(
    (
        SELECT id
        FROM companies
        WHERE id NOT IN
        (
            SELECT companyId
            FROM crawlLog
        )
        LIMIT 1
    ),
    now()
)

This could only be done reliably with a transaction and some locking. If you're on InnoDB, at least it'd only be a row-level lock and not a whole table lock. — Marc B, Jul 08 '11 at 20:52
@Tocco - you are the man!!! That did it. This is a great trick. — T. Brian Jones, Jul 08 '11 at 20:59

score 3 · Answer 1 · answered Jul 08 '11 at 21:06

Why use a Subselect? INSERT INTO ... SELECT exists:

INSERT INTO crawlLog (companyId, timeStartCrawling)
SELECT companies.id, NOW()
FROM companies
LEFT OUTER JOIN crawlLog
ON companies.id = crawlLog.companyId
WHERE crawlLog.companyId IS NULL
LIMIT 1

And that way it should not complain about using a table both in the INSERT and SELECT part

score 1 · Answer 2 · edited May 23 '17 at 12:10

You can't update rows which you are querying. There is a way to force MySQL to use a temporary table implicitly:

INSERT INTO crawlLog (companyId, timeStartCrawling)
VALUES
(
    SELECT id, when FROM
    (
    SELECT companies.id AS id, now() AS when FROM companies
    LEFT OUTER JOIN crawlLog
    ON companies.id = crawlLog.companyId
    WHERE crawlLog.companyId IS NULL
    LIMIT 1
    )
)

T. Brian Jones · Accepted Answer · 2011-07-08T21:12:10.470

1

This works and seems like the simplest solution:

Using the simpler of the two statements in my question, I created an alias for the inner crawlLog table as suggested by @Tocco in the comments, and then removed the necessary encapsulation in VALUES().

INSERT INTO crawlLog (companyId, timeStartCrawling)
SELECT id, now()
FROM companies
WHERE id NOT IN
(
    SELECT companyId
    FROM crawlLog AS crawlLogAlias
)
LIMIT 1

edited Jul 08 '11 at 21:12

answered Jul 08 '11 at 21:01

T. Brian Jones

13,002
25
78
117

1

Not sure if it's a problem or not, but putting a SELECT inside of VALUES seems odd (unnecessary). – Chains Jul 08 '11 at 21:06

score 0 · Answer 4 · answered Jul 08 '11 at 20:59

0

Do the select into a temp table, then insert selecting from the temp table. You can't insert into a table and select from it in the same statement, so use a temp table and two statements.

answered Jul 08 '11 at 20:59

antlersoft

14,636
4
35
55

How do I rewrite this MySQL query so it doesn't throw this error: You can't specify target table 'crawlLog' for update in FROM clause?

4 Answers4

Linked