Apache Hive Driver issue using CDH3

tableaucommunity@tableau.com — Wed, 09 May 2012 15:24:55 GMT

Hello,

We just had an issue with a machine where we tried to install the CDH3 ODBC driver. See the attached image... Basically, we installed the driver (see the directory in the background), however it is not showing up in the ODBC source administrator. This is a Windows 7 box that we were able to produce the behavior on. Anyone come across a fix for this?

Thanks,

Jonathan

Wildcard limitation?

tableaucommunity@tableau.com — Wed, 04 Jan 2012 01:46:21 GMT

Hello!

We've been using Tableau with great success with some awesome workbooks the last 1.5 years, but I just ran into something I can't seem to get around. Wildcard filters seem to work for contains queries, but is there a way use true wildcarding / regex for a filter?

IE: I have a parameter that I use for a bool check on a filter to determine which data to display.

For the simplicity of explanation, suppose this parameter is looking at a country name field (it isn't, but this exposes the challenge) and there are many different variants of the name Morocco. In the dataset, suppose we have:

Maracco

Morocco

Moracco

Marocco

Moracca

I need to be able to search on M*R*CC* and have it filter to show all the variants above, but nothing else... Our situation is actually far more magnified than this simple country example (we actually have on the order of 100+ variants per search) and the running entity resolution beforehand won't work either (what we have aren't mispellings and are intended to be this way... instead think an encoded GUID where each digit position in the guid is meaningful) as we actually want to pull up all the original representations (as they each mean something different)... and we can't just break out the digit positions into their own dimensions - we can't do that either as the digit positions vary :) (*sigh*)... Fortunately, the encoding is such that we can use vowels and non-vowels much like the Morocco search above, so if we can get that to work, we're in good shape. If we have to, I guess we can pre-process everything beforehand and make whole new column mapping all variants to all possible searches, but that would be bordering on insane levels of denormalization that we would rather not explore if we can just get the wildcards above to work.

If there is a way to do this, how would I then wire that up through the condition filter I have set up for that parameter? IE: right now, the best I can do is Contains(UPPER([Field1]), UPPER([Parameter1]))... is there an equivalent to Wildcard(UPPER([Field1]),UPPER([Parameter1]), '*') or something?

Also, a completely separate question to post... any word on a 64 bit thick client version being released? We have some extremely large workbooks with ~100GB of data that work great, but always run out of RAM on the 4GB 32 bit limit (and we've got well over 100GB RAM)... I know we have to be careful not to draw too many points on the screen for visualization (else the visualization becomes a mess and is worthless), but we've pretty well solved that display problem through careful filtering, pre-computing all data before Tableau, and other techniques to aim the gun away from foot... now we just need the "running low on memory crashes" to go away - so I was just curious if it was in the works :)

Tableau rocks - thanks!

-Jonathan

Content in Tableau Community Forums

Apache Hive Driver issue using CDH3

Wildcard limitation?