Обсуждение: Multi insert statement and getUpdateCount().

Hi,
  I have been testing the current version of pgjdbc master branch. I
added a test case to a branch in my repo [1] to check the behaviour when
a multi insert statement adds two rows to a table. I found the return
value when calling .getUpdateCount() does not equal the number of rows
affected.

  My question is what should the return value for .getUpdateCount()
represent. Is it the number of statements executed or rows put in the
database ?

Regards,
Jeremy

[1]
https://github.com/whitingjr/pgjdbc/tree/updatecount-mismatches-updatecount

--
Jeremy Whiting
Senior Software Engineer, JBoss Performance Team
Red Hat

------------------------------------------------------------
Registered Address: RED HAT UK LIMITED, 64 Baker Street, 4th Floor, Paddington. London. United Kingdom W1U 7DF
Registered in UK and Wales under Company Registration No. 3798903  Directors: Michael Cunningham (US), Charles Peters
(US),Matt Parson (US) and Michael O'Neill(Ireland) 

Jeremy Whiting wrote:
> Hi,
>   I have been testing the current version of pgjdbc master branch. I added a
> test case to a branch in my repo [1] to check the behaviour when a multi insert
> statement adds two rows to a table. I found the return value when calling
> .getUpdateCount() does not equal the number of rows affected.
>
>   My question is what should the return value for .getUpdateCount() represent.
> Is it the number of statements executed or rows put in the database ?
>
> Regards,
> Jeremy
>
> [1] https://github.com/whitingjr/pgjdbc/tree/updatecount-mismatches-updatecount
>

My quick test shows it seems to be represent the the rows effected. That is
using a single statement with multiple insert values. (3)

INSERT INTO "public"."key_table4" ("avatar_id", "user_id", "bing_id") VALUES
(3, 21, 20),
(5, 17, 31),
(8, 5, 7);

If you are using single insert statements I would assume it will always return
(1).

According to the Java API

Retrieves the current result as an update count; if the result is a ResultSet
object or there are no more results, -1 is returned. This method should be
called only once per result.

Seems correct, rows effected. Using postgresql-9.4-1200.jdbc4.jar, not current
repository though.

danap.

On 26/03/15 19:41, dmp wrote:
> Jeremy Whiting wrote:
>> Hi,
>>   I have been testing the current version of pgjdbc master branch. I
>> added a
>> test case to a branch in my repo [1] to check the behaviour when a
>> multi insert
>> statement adds two rows to a table. I found the return value when
>> calling
>> .getUpdateCount() does not equal the number of rows affected.
>>
>>   My question is what should the return value for .getUpdateCount()
>> represent.
>> Is it the number of statements executed or rows put in the database ?
>>
>> Regards,
>> Jeremy
>>
>> [1]
>> https://github.com/whitingjr/pgjdbc/tree/updatecount-mismatches-updatecount
>>
>
> My quick test shows it seems to be represent the the rows effected.
> That is
> using a single statement with multiple insert values. (3)
>
> INSERT INTO "public"."key_table4" ("avatar_id", "user_id", "bing_id")
> VALUES
> (3, 21, 20),
> (5, 17, 31),
> (8, 5, 7);
>
> If you are using single insert statements I would assume it will
> always return
> (1).
>
> According to the Java API
>
> Retrieves the current result as an update count; if the result is a
> ResultSet object or there are no more results, -1 is returned. This
> method should be called only once per result.
>
> Seems correct, rows effected. Using postgresql-9.4-1200.jdbc4.jar, not
> current
> repository though.
>
> danap.
>
>
  Thanks for your reply. It prompted me to do some more testing and
stepping through the code.

  I found the return value for your example will be
-1 == .getUpdateCount()

  whereas the int array return value for the method call .executeBatch()
int[] == {3}

  Thanks again for helping me to understand this better.

Jeremy