US and Worldwide: +1 (866) 660-7555

Search Forums:

Type: Posts; User: mcampbell; Keyword(s):

Page 1 of 16 1 2 3 4

Search Forums: Search took 0.12 seconds.

  1. The NullPointer you're getting is inside of saiku...

    The NullPointer you're getting is inside of saiku code, so it would be good to ask on the saiku forum if you haven't already.

    Are there any other errors logged?
  2. Can you implement this as a UserDefinedFunction? ...

    Can you implement this as a UserDefinedFunction? That'd allow you to plug in your function without hacking anything.
  3. Replies
    7
    Views
    16,891

    Hi Laura, Schema workbench is the main tool...

    Hi Laura,

    Schema workbench is the main tool used for defining schemas:
    http://help.pentaho.com/Documentation/6.0/0N0/020/070/020

    This link gives more details about schema construction:...
  4. No, Excel's not an option. An alternative to...

    No, Excel's not an option. An alternative to workbench is to use Pentaho's Data Source Wizard, which can auto-generate simple schemas.
  5. Have you looked at the SQL/MDX debug logging...

    Have you looked at the SQL/MDX debug logging during execution? That may give you some idea of why the performance is so bad, and why it differs from 3.3.
  6. Replies
    2
    Views
    258

    1) Yes, you can have any number of aggregate...

    1) Yes, you can have any number of aggregate tables. The desire is to create aggregate tables which hit the majority of attribute rollups that will be requested by your workloads. So if you have...
  7. Replies
    3
    Views
    281

    The best place for Mondrian 4 doc is Julian's...

    The best place for Mondrian 4 doc is Julian's book: https://www.manning.com/books/mondrian-in-action

    Other than that, your best bet is to use the M4 Foodmart.xml schema as an example: ...
  8. Replies
    1
    Views
    149

    So you want a measure to be null for certain...

    So you want a measure to be null for certain members?

    You can define a calculated measure that uses a conditional. Something like

    iif( MyDimension.CurrentMember IS MyDimension.MyMember,...
  9. Replies
    1
    Views
    171

    This is probably a better question for the saiku...

    This is probably a better question for the saiku forum: https://groups.google.com/a/saiku.meteorite.bi/forum/#!forum/user
  10. There are some inconsistencies in what Mondrian's...

    There are some inconsistencies in what Mondrian's XMLA issues and what AdomdClient expects.

    http://jira.pentaho.com/browse/MONDRIAN-1836
  11. Replies
    1
    Views
    105

    Are you using a parent-child hierarchy? Sounds...

    Are you using a parent-child hierarchy? Sounds like not. P-C often fits nicely with Employee type dimensions, where each employee may manage some number of employees, and you need to do things like...
  12. The easiest way to test out calculated members is...

    The easiest way to test out calculated members is by running them in MDX queries, including a WITH block at the start of your query. E.g.

    WITH MEMBER measures.foo as...
  13. Replies
    1
    Views
    201

    CrossJoin(set1, set2) should be functionally...

    CrossJoin(set1, set2)

    should be functionally equivalent to

    set1 * set2

    If you have a case with SampleData that gives different results that could be a bug. Can you pass on repro steps?
    ...
  14. Replies
    4
    Views
    350

    The answer depends a lot on the MDX, schema and...

    The answer depends a lot on the MDX, schema and data.

    Typically when I'm looking to identify the cause of poor performance I start out by turning on MDX and SQL debug logging (in log4j.xml),...
  15. CurrentDateMember might be useful to you. Using...

    CurrentDateMember might be useful to you. Using that with the Tail function can give you the set of the last 30 days.
    Checkout out Diethard's blog post on it: ...
  16. Replies
    4
    Views
    350

    The distributed cache holds cell data (segments)....

    The distributed cache holds cell data (segments). Being able to pull segments out of the cache rather than executing SQL can reduce a good chunk of execution time, but other parts of MDX evaluation...
  17. Replies
    3
    Views
    281

    Your link points to incomplete documentation for...

    Your link points to incomplete documentation for Mondrian 4.

    Mondrian 3's schema documentation can be found here: http://mondrian.pentaho.com/documentation/
  18. You need to drop your jar in the classpath. ...

    You need to drop your jar in the classpath. biserver/tomcat/webapps/pentaho/WEB-INF/lib should work.
  19. Replies
    2
    Views
    204

    One option is to use a member formatter: ...

    One option is to use a member formatter: http://mondrian.pentaho.com/documentation/schema.php#Member_formatter

    Alternatively you could place the si/no values in a char field and reference that...
  20. It looks like you're crossjoining many single...

    It looks like you're crossjoining many single members. If each argument in a crossjoin is a single member or tuple then the result is a tuple set of size 1, which is more easily represented as just...
  21. Replies
    5
    Views
    2,067

    LookupCube() isn't implemented in Mondrian. ...

    LookupCube() isn't implemented in Mondrian.

    Most places where you might want to use lookupCube are better handled by bringing the cubes together in a virtualcube or measure group. (IIRC this is...
  22. Replies
    4
    Views
    312

    For your 1st option, I'd start out with a sanity...

    For your 1st option, I'd start out with a sanity check-- set your caption SQL expression to something like a fixed literal string value (say 'FOO'). That way you can confirm the caption expression...
  23. Replies
    4
    Views
    312

    There are a couple ways to control how a member...

    There are a couple ways to control how a member will be displayed:
    1) use a captionColumn defined on the level, pointed to a column with the desired caption to be associated with each member. ...
  24. Replies
    4
    Views
    407

    Is employeeStatus a degenerate dimension? I...

    Is employeeStatus a degenerate dimension? I wonder if it's trying to find a join path between your agg table and the fact table to handle degenerates.

    You could try setting employeeStatus to...
  25. Can you post the MDX and some of these multiple...

    Can you post the MDX and some of these multiple queries from your log file?
  26. Replies
    1
    Views
    219

    You might be best off asking on the saiku forum.

    You might be best off asking on the saiku forum.
  27. Looks like you're using integrated security. ...

    Looks like you're using integrated security. Does your user have permission on the database? Are you able to connect from other clients using that account?
  28. Replies
    3
    Views
    252

    Can you post your MDX? A compound slicer...

    Can you post your MDX?

    A compound slicer should effectively be the same as wrapping the set of members in an Aggregate(), giving the rolled up value of the set.
  29. There are a couple issues with your query- ...

    There are a couple issues with your query- Drillthrough should select a single cell from a cube. You have sets of members on COLUMNS and PAGES, so the DRILLTHROUGH is incorrectly being applied to...
  30. Replies
    3
    Views
    252

    I think Saiku has a "Filter" area where you can...

    I think Saiku has a "Filter" area where you can drag a level, then select one or more members to include in the filter. In MDX that amounts to a compound slicer, e.g.

    WHERE
    {...
  31. Use of the asterisk crossjoin operator can cause...

    Use of the asterisk crossjoin operator can cause performance issues because it is not flagged for push-down optimization (whereas NonEmptyCrossJoin or CrossJoin would be). Since Saiku tends to use...
  32. There may be issues loading metadata via ODBC. ...

    There may be issues loading metadata via ODBC. Can you use a JDBC connection?
  33. The Sum() function takes a set as its first...

    The Sum() function takes a set as its first argument, not the numeric value you want to sum. You could do something like

    SUM( Descendants( Geography.CurrentMember, Geography.Zipcode),...
  34. There are limitations, e.g. only some fairly...

    There are limitations, e.g. only some fairly basic calculations are allowed as the 3rd argument. How is your Measures.[VAT] defined?

    Also, the set expression in the first expression has to be...
  35. Replies
    3
    Views
    244

    Can you attach your logs?

    Can you attach your logs?
  36. Replies
    2
    Views
    271

    Hi vjsp90, The "system properties" log message...

    Hi vjsp90,

    The "system properties" log message actually refers to system properties (i.e. retrieved from System.getProperties()), not the properties loaded from the mondrian.properties file.
    ...
  37. Replies
    2
    Views
    272

    There isn't a well defined, canonical spec for...

    There isn't a well defined, canonical spec for MDX in quite the same way as SQL. Mondrian looks to SSAS as the de facto standard for MDX and in most features should follow the same semantics. There...
  38. There's a trick you can use to get fact level...

    There's a trick you can use to get fact level stddev: For whichever additive base measure you'd like to compute, include an additional base measure that is its square value. Along with that,...
  39. Replies
    1
    Views
    256

    I don't believe using that xaction will reliably...

    I don't believe using that xaction will reliably clear cache post 5.0. You should be able to make a GET request to refresh/mondrianSchemaCache, though, like:

    ...
  40. Hmmm. I tried reproducing with the query below...

    Hmmm. I tried reproducing with the query below against Foodmart and never saw passes exceed 1. Anything unusual about any of the referenced dimensions? Is Employee P-C?



    WITH
    set [Top0]...
  41. The way I load Foodmart into different database...

    The way I load Foodmart into different database types is to use the MondrianFoodmartLoader, which is compiled in Mondrian's test classes. I.e. clone the git repo...
  42. Hi Eric, The way to "link" two sets of tuples...

    Hi Eric,

    The way to "link" two sets of tuples is to crossjoin them. E.g. the following would compute the [total quantity] stddev at the granularity of Product.Order*Traders.Order:

    Stddev( ...
  43. Take a look at the Mondrian property...

    Take a look at the Mondrian property mondrian.rolap.evaluate.MaxEvalDepth: http://mondrian.pentaho.com/documentation/configuration.php

    Complex calculations sometimes need multiple evaluation...
  44. Trying to treat an attribute as a measure is...

    Trying to treat an attribute as a measure is rarely the right choice for lots of reasons.

    Rather than try to force two granularities into a single fact table, you could explore using 2 separate...
  45. Yes, that's a problem. Attempting to do...

    Yes, that's a problem. Attempting to do statistical functions on such big sets is a huge performance challenge with Mondrian.

    Not a particularly helpful answer, but the one time I needed to...
  46. Do you see these repeated "getMemberChildren"...

    Do you see these repeated "getMemberChildren" queries if you use your UDF with a Foodmart query? If so let me know and I'll see if I'm able to reproduce.
  47. Many functions like Median, Avg, StdDev, etc. can...

    Many functions like Median, Avg, StdDev, etc. can result in poor performance when their set arguments are largish, and the calculation is evaluated repeatedly. For example, if you have a large...
  48. Yes, agg tables can still work with aggregate...

    Yes, agg tables can still work with aggregate tables. The base measure components of a calculation can be loaded from an agg table. The query calculations will still be done in Mondrian (i.e. the...
  49. getMemberChildren calls are supposed to be...

    getMemberChildren calls are supposed to be cached, so I'm surprised you're seeing the query repeated. Have you disabled any flavors of caching in your properties file?
  50. Why not make batch_qty a measure? BTW- the...

    Why not make batch_qty a measure?

    BTW- the reason your member causes that error is that you're dividing a numeric measure by a dimension level [batch].[batch].[batch_qty]. Mondrian can't make...
Results 1 to 50 of 775
Page 1 of 16 1 2 3 4