Should I use separate databases

[micstr]

Hi,

I see, that some people write they have separate databases for their own game and for datamining. Is it common, or in any way recommended? If so, what are the pros and cons?

[WhiteRider]

If you import a very large volume of hands it can help performance to split your databases, but for normal play it shouldn't be necessary.

[micstr]

Ok, let me put it this way, I import about 120k hands everyday. My database is about 8 milion hands now, but my own hands are 150k.

I'm not sure if I should purge my database every now and then? What does purge do exactly? How should I do it not to lose some important data?

[WhiteRider]

In that case separate databases will likely work better for you.
You may prefer to start a new database periodically (just keep all of your own hands in a single database) - purging removes hands from the database permanently, and if you're removing a large number of files it can take a significant time so you'd be better dropping an old database and starting a new one. You can configure the HUD to use the last 2 or 3 databases for example.

[COKE_MAN]

He is asking what happens when he purges hands. Does this change what is shown for a particular stat since data points are missing? Does the cache account for this? Is this only an issue when the cache is rebuilt? etc.

[WhiteRider]

Any hands purged are removed from the database completely. Stats will not be used from those hands which were removed. The cache is updated automatically after purging - you no longer need to do any housekeeping manually.

[jspill]

I was about to switch to PT4 from HEM1, and have now found out the above issue - that purging removes hands completely.

On HEM1 you can purge HH files, drastically reducing the size of your database, but still have all the stats from millions of hands on the HUD. This was perfect for importing 200k+ hands per day.

If this isn't possible on PT4 there's just no way I can give up HEM1, even though I'd like to switch for other reasons. I'm amazed it isn't possible.

Could you please outline how I could use multiple databases as a workaround/link me to some threads? How does that help? Wouldn't one database still be huge? I don't have space for a 300GB database. My current HEM1 database has over 10 million hands, and is just 15GB.

[WhiteRider]

This was written for PT3, but you may find it useful: Tutorial: Managing Multiple Databases

[micstr]

How is it possible, that when I import HHs from PT3 /processed directory into PT4, they get kind of messed up? I import them into a new database, I get about 150k hands imported, but when I configure aliases, I only get about 70k hands. I import manually from disk. Maybe I should do it differently?

Or is there a way to extract only mine hands from the old database and make a new one from them?

I'm trying to make separate databases and I think the easiest way would be to start a fresh one only for played hands, and keep the huge old one for datamining and gradually purging the limits I dont play anymore and maybe older than - lets say - 3 months. Maybe in some time start a new one to keep things tidy.

[kraada]

You can export just your hands via Database --> Database Management --> Export. Select your name and highlight all sessions then choose export to export just those sessions. Note: you can highlight multiple rows quickly by clicking on the first row, pressing and holding shift, then clicking on the last row.