[Mondrian] (on latest builds of) Mondrian 3.4: Mondrian Error:Internal error: Reading cardinality for column

Julian Hyde jhyde at pentaho.com
Fri Apr 13 13:24:47 EDT 2012


On Apr 13, 2012, at 7:06 AM, Pedro Alves wrote:

> Ok, this is not as bad as I initially thought.
> 
> - For some reason, on our build, the code that we have in 
> mondrian-3.4.0-SNAPSHOT.jar is, in fact, code for master. We can see 
> there classes that belong to the commit in 
> https://github.com/pentaho/mondrian/commit/3129b4ee047ad251df472c483975c3beb4b2da81#diff-0
> 
> - So while there's definitely a bug that's affecting (at least) vertica, 
> it's not in 3.4

You probably already know this. But I'll state the obvious.

The change is caused by by my edit to add a StatisticsProvider SPI. I intentionally committed this to master -- aka the 3.5 branch :) -- because it was too risky to put into 3.4.

http://github.com/pentaho/mondrian/commit/3129b4ee047ad251df472c483975c3beb4b2da81

There is a suspicion -- Luc can confirm -- that this change has somehow leaked into the 3.4 branch. That is bad news because 3.4 is supposed to be super-stable pending the release.

Meanwhile, back on the master branch, I will work on a fix for postgres.

Julian



More information about the Mondrian mailing list