cube - SSAS tabluar mode processing fails with "a lot of rows" -
cube - SSAS tabluar mode processing fails with "a lot of rows" -
i have ssas tabular mode cube reads info actian matrix database using odbc. project processes fine when i'm using info set 1 1000000 rows when seek utilize bigger 1 (300 1000000 rows), process runs around 15 minutes , fails message:
the operation failed because source database not exist, source table not exist, or because not have access info source. more details: ole db or odbc error: [paraccel][odbc driver][padb]57014:error: query (25459) cancelled on user's request detail: query (25459) cancelled on user's request ; 57014. error occurred while processing partition 'xxxx' in table 'yyyy'. current operation cancelled because operation in transaction failed. the message says database doesn't exist doesn't create sense because works fine on first case (and difference "where clause" limit number of rows)
i'm using server has 96 gb of free ran , can see memory beingness consumed while "processing" process running. when consumed, runs few axtra seconds , fails. also, know fact 300 1000000 row dataset exported csv file has 36 gb on raw format, should fit total in memory without compression.
i can guarantee query works fine on own on source database "query (25459) cancelled on user's request" message doesn't create much sense.
does have thought on may going on?
memory consumption on derivative of input rows (the resulting cube) cannot estimated on byte size of input. function of cartesian graph product of distinct values of cube dimensions.
if building cube 2 input rows on 2 dimensions , 2 measurements:
state|city|population --------------------- ny|new york|8406000 ca|los angeles|3884000 state|city|population|number of records --------------------------------------- null|null|12290000|2 ny|null|8406000|1 ny|new york|8406000|1 ca|null|3884000|1 ca|los angeles|3884000|1 null|los angeles|3884000|1 null|new york|8406000|1 you can't expect output beingness generated input info rows processed equivalent in size. if odbc driver keeps entire input in memory before lets read it, have business relationship both input , output reside in memory until cube generation complete.
this reply much clearer on subject: how calculate size of olap cube
ssas cube olap-cube tabular paraccel
Comments
Post a Comment