TM1 API string limit

[Wim Gielis]

Hello all,

While creating a TI process from inside Excel VBA (I'm almost there !), I run into an issue of I think a String limit of 1,024 characters.

For example, if we put code in the Prolog of a TI process, after 1,024 characters (vbCrLf characters added, which is 2 characters every time) we find an extra CRLF character.
In itself, this isn't too much of a problem, we can just delete the few extra CRLF characters and we're done.

But when the extra CRLF breaks a commentary line in 2 pieces, the process won't save since syntaxically there is an error.

For example, have a look at the line with "SubsetCreateByMDX" for the dimension }Processes where effectively TM1FilterByLevel has been chopped off:

Code: Select all

vCube = 'General_Process control' ;

vView_Tgt = 'ZZZ_tmp' | '_' | Timst( Now, '\Y\m\d\h\i\s' ) ;
vSubset_Tgt = 'ZZZ_tmp' | '_' | Timst( Now, '\Y\m\d\h\i\s' ) ;

# Turn off target cube log changes
CubeSetLogChanges( vCube, 0 );

# Custom settings to output information to a text file
DataSourceAsciiDelimiter = ';';
DataSourceAsciiQuoteCharacter = '';

##########
# Zero Out the Target View
##########

# Start from scratch: destroy the temporary objects
ViewDestroy( vCube, vView_Tgt );

SubsetDestroy( '}Processes', vSubset_Tgt );
SubsetDestroy( 'General_Process measures', vSubset_Tgt );

# Create a temporary view
ViewCreate( vCube, vView_Tgt );

# Limit the contents of the view
ViewExtractSkipRuleValuesSet( vCube, vView_Tgt, 1 );
ViewExtractSkipZeroesSet( vCube, vView_Tgt, 1 );
ViewExtractSkipCalcsSet( vCube, vView_Tgt, 1 );
ViewExtractSkipConsolidatedStringsSet( vCube, vView_Tgt, 1 );

# Create temporary subsets in dimensions

SubsetCreateByMDX( vSubset_Tgt, '{TM1Sort( {TM
1FilterByLevel( {TM1SubsetAll( [ }Processes ] )}, 0)}, Asc)}', '}Processes' );
SubsetCreateByMDX( vSubset_Tgt, '{TM1Sort( {TM1FilterByLevel( {TM1SubsetAll( [ General_Process measures ] )}, 0)}, Asc)}', 'General_Process measures' );

# Adding subsets to the view
ViewSubsetAssign( vCube, vView_Tgt, '}Processes', vSubset_Tgt );
ViewSubsetAssign( vCube, vView_Tgt, 'General_Process measures', vSubset_Tgt );

# Clear existing data
ViewZeroOut( vCube, vView_Tgt );

# Tidy up: destroy temporary objects
ViewDestroy( vCube, vView_Tgt );

SubsetDestroy( '}Processes', vSubset_Tgt );
SubsetDestroy( 'General_Process measures', vSubset_Tgt );

Here is the relevant portion of the code (without error handling, just the code):

Code: Select all

    '1 - PROLOG
    sqCode_Prolog = Split(sCode_Prolog, "§")
    CountOfLines = UBound(sqCode_Prolog) + 1

    ReDim PrologLines(CountOfLines - 1)

    For i = 0 To CountOfLines - 1
        PrologLines(i) = sqCode_Prolog(i)
    Next

    ReDim lArray(CountOfLines)
    hArray = TM1ValArray(hArrayPool, lArray(), CountOfLines)

    For m = 0 To UBound(PrologLines)
        TM1ValArraySet hArray, TM1ValString(hsPool, PrologLines(m), 0), m + 1
    Next

    lReturn6 = TM1ObjectPropertySet(hppPool, lReturn4, TM1ProcessPrologProcedure, hArray)

sCode_Prolog is a String variable with each time the character § to mark a new line.


Does this limit of 1,024 characters pertain to the function TM1ObjectPropertySet ? Or Tm1ValString ?
Any pointers towards a solution of avoiding the extra CrLf character ?

Thanks a lot !

[BrianL]

The only limit I'm aware of is that TM1ValString can only create a string with length less than 65535.

You could always iterate and output PrologLines and hArray to verify they contain the expected data.

[Wim Gielis]

Thank you Brian, I will work on this.

As I have the correct output in the Prolog tab, and the extra vbCrLf is *exactly* each time after every 1,1024 characters, it is something strange to my mind.
If we have a lof of code in a TI tab, the extra vbCrLf is put more than once, each time after 1,024 characters.
In any case, while the String variable sCode_Prolog contains the full TI code for the Prolog tab, it is chopped up and passed line after line to the functions.

Thanks again.

[BrianL]

Yeah, it's a strange issue. While I have no VBA experience, I don't see anything that looks wrong in your code. All the TM1 API calls match what I'd expect to see in the C API. On the other hand, I've seen lines longer than 1024 characters in a TI created in architect which indicates it's not some fundamental API limit...

Real curious to see how this pans out.

[Wim Gielis]

Real curious to see how this pans out.

Me too

Thanks for confirming that it will not be an obvious errer of my part

[Wim Gielis]

Duh... Error from my side.

The String variable sCode_Prolog already had vbCrLf characters to separate lines, rather than § (my separation character).
Therefore, splitting sCode_Prolog on § yielded an array with only 1 element, the very long String.
In that case, it seems that outputs with an extra vbCrLf each 1,024th character.
When I split the long string on vbCrLf, it gives me more than 100 lines in the Prolog, each line being much smaller. Then TM1 adds them fine.

Problem solved, thanks BrianL.