dramatic different in selectRows() time for no obvious reason | Ben Bimber | 2011-02-06 06:55 |
Status: Closed | ||
I'm using selectRows() to return the structure of a table. I'm using with through the LABKEY store, but this example separates it for simplicity. I'm using maxRows=0, so I only return the metadata. I am only selecting physical columns from the table, ie. no wrapped columns or using lookups to add columns. take this URL: http://localhost:8080/labkey/query/WNPRC/EHR/selectRows.view?_dc=1297002527146&schemaName=study&query.queryName=Clinical%20Remarks&query.maxRows=0&apiVersion=9.1&query.columns=lsid%2Cobjectid%2Cqcstate%2Cremark%2Cid%2Cdate%2Cproject%2Caccount%2Ctaskid I select a bunch of the columns. It runs in a second or two. Then this one: http://localhost:8080/labkey/query/WNPRC/EHR/selectRows.view?_dc=1297002527146&schemaName=study&query.queryName=Clinical%20Remarks&query.maxRows=0&apiVersion=9.1&query.columns=parentid The only different is the columns selected. Simply asking for the parentId column takes from 20-60 secs. ParentId is a field in this dataset. It is a lookup to studyData, which cant be fast, but I dont get why maxRows=0 would incur a problem here, since we're not asking for any data. Could it have something to do with labkey resolving the lookup anyway to try to find the displayValue of that field? While it's nice that labkey tries to be friendly and automatically show the displayField of every lookup, it seems like some manner to turn this off (say if the displayField on the target equals the lookup field??) would be a not insignificant perf boost. |
||