• Article Type: General
• Product: Aleph
• Product Version: 20

Description:
Because it has been a couple of years since we implemented publishing, we thought it time to do some maintenance surrounding publishing within Aleph.

So, I first decided on our test server to drop and recreate the z00p table completely. I just completed that process yesterday.

What I found was that the size of the resulting z00p table was only about 68% of the size of the current z00p table. I dropped from about 34.3M records down to 23.4 M, which seems like a very large decrease.

Is this something I should have expected?

I ran this using p_publish_04 using this type of command line:

csh -f /exlibris/aleph/a20_1/aleph/proc/p_publish_04 ABC01,PRIMO-ABC,001000000,001250000,5,

Resolution:
I think this has to do with suppressed records. Comparing the following SQL on production to test, though there are cases where there is no z00p on production, there are many additional ones on test which lack the z00p. I find all of these have "902 $$aSuppress display". (Suppressed records should be omitted from the ABC01PUB base.) Note: "z13 where z13_title is not null" is used instead of z00 so that (logically) deleted records are omitted from the results.


production:

abc01@MALC20> select count(*) from z13 where z13_rec_key in (select z13_rec_key from z13 where z13_title is not null minus select z00p_doc_number from pwd50.z00p where z00p_set = 'PRIMO-ABC');
**** Hit return to continue ****

COUNT(*)
----------
66359


test:

abc01@MALC20> select count(*) from z13 where z13_rec_key in (select z13_rec_key from z13 where z13_title is not null minus select z00p_doc_number from pwd50.z00p where z00p_set = 'PRIMO-ABC');
**** Hit return to continue ****

COUNT(*)
----------
1097864


You can use this SQL to get individual record keys:

abc01@MALC20> select z13_rec_key from z13 where z13_title is not null minus select z00p_doc_number from pwd50.z00p where z00p_set = 'PRIMO-ABC';


test is correct in not creating the z00p when the record is suppressed. I think the result of your recent p_publish_04's is correct.

• Article last edited: 10/8/2013