gregs

Show:

What's happening?

  • Comment: Jailer importing CLOB's incorrectly.

    Thanks Ralf, that appears to have fixed the bug.

    2009-10-02 14:30:55 UTC in Jailer

  • Comment: Jailer importing CLOB's incorrectly.

    Ralf, I'm using 3.1.7 and replaced the jailer.jar file with the 3.1.8.beta version. I get an error message saying "Invalid property 'identityInserts' of bean class [net.sf.jailer.configuration]: No property 'identityInserts' found (InvalidPropertyException). Is there something else I need to do to use the new jar? Thanks, Greg.

    2009-10-02 13:59:05 UTC in Jailer

  • Comment: Jailer importing CLOB's incorrectly.

    To clarify, the problem looks to be being caused by the import process missing the space character in the Jailer import file. The position of the space in the database has no bearing on the issue. It seems that when Jailer writes a CLOB to the import file it starts a new line every 100 characters. If a space character happens to be the 100th character and thus the last character on the line...

    2009-10-02 12:16:24 UTC in Jailer

  • Comment: Jailer importing CLOB's incorrectly.

    If you look at lines 1263 and 1302 in the ADAPTER_RDL.sql file you'll see these two lines have a space as the last character in the line. If you look at the imported_clob file, you'll see the space has been ommited during the import on lines 9 and 28. If you compare lines 9 and 28 of the exported_clob file withe the imported_clob file you'll see the difference.

    2009-10-02 11:34:35 UTC in Jailer

  • Jailer importing CLOB's incorrectly.

    When importing a multi line CLOB that has a "space" character at the end of a line, the importer tool ignores the space and concatenates the first character of the next line to the character immediately before the "space" character. This causes seperate words to be joined together.

    2009-10-02 11:27:44 UTC in Jailer

  • Followup: RE: SQL import performance.

    Hi Ralf, The enhancement works well. I extracted the same data as before the fix with the max rows parameter set at 750 and it shaved about 80% off the load time! Thanks, Greg.

    2009-07-07 13:21:28 UTC in Jailer

  • Followup: RE: SQL import performance.

    Hi Ralf, Single row inserts for LOB's won't be a problem. I see single row inserts for LOB's as preferable, as the user defined max rows per statement would probably have to be set at a low number to accomodate LOB's, thereby forcing the non-LOB data to use the same low number for max rows per statement. Thanks for your efforts, Greg.

    2009-07-06 10:35:46 UTC in Jailer

  • Followup: RE: SQL import performance.

    Hi Ralf, The method selecting from dual with the union command gave the best results. The "Insert All" method throws an ora-24335 exception if you try too many lines, whereas the dual/union method just gets slow if you throw too many rows at it. I loaded 1000 rows in 1s and 5000 rows in 22s with the dual/union method, so I guess it depends on the database and server memory as to...

    2009-07-03 16:25:47 UTC in Jailer

  • Followup: RE: SQL import performance.

    Ralf, I'm using Oracle, which I guess doesn't support multi-row inserts? I will give DbUnit a shot. Thanks, Greg.

    2009-07-02 12:37:41 UTC in Jailer

  • SQL import performance.

    Hi, I've been using Jailer to extract data subsets from a very large production database and it works well. However, my subsets are around 4-6 million rows in size due to the large amount of static/standing/config data required in my test database and are taking a very long time to load. It seems that jailer is inserting and commiting one row at a time to the database, which is fine for...

    2009-07-01 14:19:26 UTC in Jailer

About Me

  • 2009-07-01 (4 months ago)
  • 2549928
  • whangarei (My Site)
  • gregs

Send me a message