Beta 15 Released

[ffrllc]

Afterwards, PLEASE do a cluster, vacuum and then analyze.

Does it matter in what order the cluster, vacuum, and analyze operations are done?

Would be nice to have an optional menu command that will do all three in one go as PT2 has.

I was thinking the same thing.

[Vytenis]

I have one database and it has about 50k hands. Is it normal that update took just 4 seconds ?

[jed3y3]

I have one database and it has about 50k hands. Is it normal that update took just 4 seconds ?

Hmmmm i got like 700K hands and mine also took 4 seconds??????

Oh well il cluster analyse etc and see what happens

Thanx for ur reply b4 buford

[PTuser670]

Import of observed PokerStars hands (max 30 per table), Hud on observed PokerStars tables

how is this not datamining? Did you okay these changes with reps from pokerstars ?

i thought stars allowed this in general as of fairly recently with the idea of it being used for table selection instead of datamining.

[WhiteRider]

Import of observed PokerStars hands (max 30 per table), Hud on observed PokerStars tables

how is this not datamining? Did you okay these changes with reps from pokerstars ?

Yes, this has been agreed with PokerStars.

When having preferred seating the HUD does not place correct until Hero's stats show up. Thus if I am seating down in middle position waiting for BB they will all be wrongly placed until I have played the first hand so my stats are read.

I believe this has to do with that the table rotates to my preferred seat as soon as I sit down, but the HUD doesn't notice it until I have played my first hand.

Yes, that's right - until the hand history file is read, PT3 doesn't know about your new seat position.

Does it matter in what order the cluster, vacuum, and analyze operations are done?

AFAIK, Cluster includes a Vacuum, do this first, then Analyse.

So no reimport is necessary to fix the import errors? PT3 takes care of that at its first startup?

PT3 will do what it needs to do to the database at first start up, but it will not fix errors with hands which were previously imported.
So for instance if you previously imported a hand where the SB completing wasn't counted as a limp, then this will not be corrected and you will need to re-import.

[pr1mo]

so if i import and have both cluster and analyze checked it'll do one after the other properly as intended? and i dont ever have to vacuum?

[WhiteRider]

That's how I understand it, yes - but hopefully one of the postgres experts will confirm this for me.

[driemekasten]

Afterwards, PLEASE do a cluster, vacuum and then analyze.

Would be nice to have an optional menu command that will do all three in one go as PT2 has.

I was thinking the same thing.

Me too..
And it would be nice to know the official way to do cluster/vaccum/analyze by script. I'm running postgresql on my Linux-Server and it would make sense to let the cluster/vaccum/analyze run nightly. Maybe one of the postgresql-experts will tell us.

[fiammaz]

I post this screenshot here so anyone can see it and check it.
http://img134.imageshack.us/my.php?image=lolpt3av8.jpg

These are the same stats on PT2 and PT3, I imported the SAME hands, and filter to only one site (pokerstars).
There are a difference of players because i played in a freeroll and i can't purge tournament in PT3, anyway it doesn't matter because it haven't much effect on the overall stats. The yellow circle is the strange data nice difference betwen PT3 and PT2... the blu circle is another strange stat, where does that number come from? Also the number of the hands are different even if I imported the same hands. As you can see the difference is enormous. How can we'll be able to analyze our game if the stats are so strange? Now the question is: Are the other stats right?

[keggler]

If you are going to CLUSTER then the VACUUM is pretty much redundant. The ANALYZE is still useful but should be done after CLUSTER since its physical-order-correlation stats will be quite wrong if done beforehand. In other words there is only one sane way to do this and it is:

CLUSTER;
ANALYZE

You could possibly make a case for

CLUSTER;
VACUUM;
ANALYZE

The VACUUM won't do anything useful in terms of reclaiming space (there
being none to reclaim just after a CLUSTER) but it would ensure that all
rows in the table are marked as committed-good, rather than leaving that
work to be done by the first transaction that happens to hit each row.


susan