*** zxiiro has quit IRC | 00:45 | |
*** Alex| has quit IRC | 01:13 | |
*** clach04 has joined #ingres | 01:28 | |
*** cytrinox has quit IRC | 02:31 | |
*** cytrinox has joined #ingres | 02:32 | |
*** PaulM05 has joined #ingres | 03:49 | |
*** Dejan has joined #ingres | 04:44 | |
Dejan | hello everyone | 04:44 |
---|---|---|
Dejan | :) | 04:44 |
PaulM05 | hi Dejan | 04:44 |
*** dyki has joined #ingres | 04:46 | |
*** zxiiro has joined #ingres | 05:07 | |
*** ChanServ sets mode: +v zxiiro | 05:07 | |
*** PaulM05 is now known as PaulM05_lunch | 05:08 | |
*** pboro has quit IRC | 05:16 | |
*** pboro has joined #ingres | 05:18 | |
*** PaulM05_lunch is now known as PaulM05 | 05:40 | |
*** mull has joined #ingres | 06:07 | |
*** Dejan has quit IRC | 06:26 | |
*** DerMeister has joined #ingres | 06:32 | |
*** Dejan has joined #ingres | 06:49 | |
Dejan | hello | 06:50 |
PaulM05 | hello again Dejan | 06:50 |
Dejan | how's going, Paul? :) | 06:50 |
PaulM05 | not bad - could do with being cooler - or better aircon :) | 06:51 |
Dejan | yeah, it is hot outside... | 06:52 |
*** DarylM has joined #ingres | 07:03 | |
*** rossand has joined #ingres | 07:47 | |
*** ChanServ sets mode: +o rossand | 07:47 | |
*** FrankW_ has joined #ingres | 08:46 | |
*** FrankW has quit IRC | 08:46 | |
Dejan | is it possible to specify that PK index of the table should use storage structure, instead of making a separate index table? | 09:15 |
Dejan | (in the create table statement) | 09:15 |
PaulM05 | yes | 09:15 |
Dejan | thanks, looks like it is the time to read create table docs | 09:17 |
PaulM05 | looks like you need to put "WITH AUTOSTRUCT" - I thought you could put "WITH INDEX=BASE TABLE STRUCTURE" in the PRIMARY KEY clause but it doesn't seem to allow that. | 09:26 |
PaulM05 | there's also a CBF param table_auto_structure which does the same thing when set to ON | 09:27 |
Dejan | PaulM05, so ... CREATE TABLE mytable(mytable_id INTEGER PRIMARY KEY WITH INDEX=BASE TABLE STRUCTURE, ...) ? | 09:32 |
PaulM05 | doesn't appear to work | 09:33 |
PaulM05 | CREATE TABLE mytable (...) with AUTOSTRUCT will use the storage structure though | 09:33 |
Dejan | all i want is to make a table where PK is going to use the structure | 09:34 |
PaulM05 | ok | 09:34 |
Dejan | i need to read docs and do some tests... | 09:34 |
Dejan | :) | 09:34 |
PaulM05 | the later will work but it's the other way around - it uses the PK to define the storage structure | 09:35 |
PaulM05 | oh and the 9.3 manuals talk about "WITH INDEX=BASE_TABLE_STRUCTURE" - note the underscores - but if you do that it'll work... and create an index called "base_table_structure" | 09:36 |
Dejan | :))))))))))) | 09:41 |
Dejan | help table says the table allows duplicates even though i wrote WITH NODUPLICATES :) | 09:43 |
Dejan | ok, HASH UNIQUE must be set... | 09:44 |
PaulM05 | or BTREE? | 09:45 |
Dejan | yes, sorry, i meant UNIQUE is needed | 09:45 |
* Dejan sees WITH NODUPLICATES as a pointless parameter | 09:45 | |
Dejan | if Ingres is going to ignore it | 09:45 |
Dejan | at least it should return an error | 09:46 |
Dejan | saying "you need unique storage structure" | 09:46 |
PaulM05 | well to be fair it's not supposed to work that way | 09:46 |
Dejan | even without unique storage structure, with noduplicates should ensure we have no duplicate tuples | 09:47 |
Dejan | right? | 09:47 |
PaulM05 | I believe so - I don't really use NODUPLICATES but I think that's the case | 09:48 |
Dejan | ok we know there are actually no duplicate rows (TIDs ensure so) | 09:48 |
Dejan | but it is confusing, and not well documented i think | 09:49 |
Dejan | i wrote a simple create table statement with noduplicates, and help table still showed that duplicate rows are allowed | 09:49 |
Dejan | it may be "help" bug | 09:49 |
Dejan | or may not | 09:49 |
PaulM05 | what was the structure of the table? | 09:50 |
Dejan | first i use default, which is heap i think | 09:50 |
Dejan | then i made it into a hash | 09:50 |
*** grantc has quit IRC | 09:51 | |
Dejan | and only when i modified it into a HASH UNIQUE ON <PK> | 09:51 |
PaulM05 | OK - so the manual says that it doesn't enforce it with heap, and that the table structure overrides it - so if you made it hash without the unique then it would allow duplicates | 09:51 |
Dejan | it shown that it allows no duplicates | 09:51 |
PaulM05 | it is a little confusing | 09:52 |
Dejan | my point exactly | 09:52 |
PaulM05 | anyway - I'm off now - have a pleasant evening | 09:58 |
*** PaulM05 has quit IRC | 09:58 | |
Dejan | you too mate | 09:58 |
*** Mud has joined #ingres | 10:00 | |
Dejan | time to go home, have fun guys | 10:40 |
*** Dejan has quit IRC | 10:40 | |
*** zxiiro has quit IRC | 11:05 | |
*** Mud has quit IRC | 11:17 | |
*** Mud has joined #ingres | 11:22 | |
*** withdefault has quit IRC | 12:12 | |
*** DerMeister has quit IRC | 12:55 | |
*** zxiiro has joined #ingres | 13:30 | |
*** ChanServ sets mode: +v zxiiro | 13:30 | |
*** Mud has quit IRC | 13:39 | |
*** zxiiro has quit IRC | 14:08 | |
*** mull has quit IRC | 14:38 | |
*** zxiiro has joined #ingres | 15:24 | |
*** ChanServ sets mode: +v zxiiro | 15:24 | |
*** rossand has quit IRC | 18:11 | |
*** cytrinox_ has joined #ingres | 19:21 | |
*** cytrinox has quit IRC | 19:24 | |
*** cytrinox_ is now known as cytrinox | 19:24 | |
*** zxiiro has quit IRC | 21:09 | |
*** cytrinox_ has joined #ingres | 22:42 | |
*** cytrinox has quit IRC | 22:45 | |
*** cytrinox_ is now known as cytrinox | 22:45 | |
*** grantc has joined #ingres | 23:47 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!