I’m trying to join 2 arrays, but I get an error message that they are not “conformant”
AFL% show(dayArticleCount_1_2_5); i,schema 0,"dayArticleCount_1_2_5<articleCount:double> [day1=15339:15726,1,0]" AFL% show(prediction_1_2_5); i,schema 0,"not empty prediction_1_2_5<multiply:double> [day=15339:15726,388,0,vect=0:0,1,0]" AFL% cross_join(dayArticleCount_1_2_5, prediction_1_2_5, day1, day); UserException in file: src/query/ops/cross_join/LogicalCrossJoin.cpp function: inferSchema line: 207 Error id: scidb::SCIDB_SE_INFER_SCHEMA::SCIDB_LE_ARRAYS_NOT_CONFORMANT Error description: Error during schema inferring. Arrays are not conformant. Failed query id: 1101434778351
My data in multiple arrays is such that some of the values in prediction_1_2_5 should be null / empty - it should be a sparse array. But that has not happened so I am trying to to only extract the values from prediction_1_2_5 for the days where the dayArticleCount_1_2_5 is zero.
I’ve tried transferring the data from prediction_1_2_5 to an array that is allowed to be empty, and I’ve tried transferring the data in dayArticleCount_1_2_5 to an array that is only non-empty, but still the cross_join fails.
Alternatively if I could do something like:
select * from prediction_1_2_5 where day in (select day1 from dayArticleCount_1_2_5 where articleCount <> 0);
that would work as well. Currently my fall back position is to export the data and carry out the analysis in python.