Opened 17 years ago
Closed 17 years ago
#316 closed defect (fixed)
SQLite: Theme function in Studio (2008) is broken
Reported by: | jbirch | Owned by: | traianstanev |
---|---|---|---|
Priority: | major | Milestone: | |
Component: | SQLite Provider | Version: | 3.3.1 |
Severity: | 3 | Keywords: | |
Cc: | External ID: |
Description
Trying to theme a SQLite table in MapGuide Studio 2008 causes Studio to fail with an exception and stack trace.
Can provide screen shot of error and source data on request.
Change History (5)
comment:1 by , 17 years ago
comment:2 by , 17 years ago
It's doing a UNIQUE call to the MapAgent's SelectAggregates function. I've tested this with a number of string columns and they all perform in the same manner.
Looks like there's some corruption of data. The last line of data always seems to contain odd characters for a unique call on a string column. Just performed the same query on the source SDF file and didn't get the last line of corrupt data. Request headers and output from one of the MapAgent calls available here:
http://www.jasonbirch.com/temp/broken_theme.zip
Data available on request.
comment:3 by , 17 years ago
Unique() is actually computed by MapGuide itself -- it simply lists all the values for the column from the provider and then does its own unique. So it is getting some garbage strings from the provider using a regular select -- which should in theory work. So it's a bug with regular select. I wonder if it has to do with null strings -- if I can't reproduce it with my data, I'll ask you to share yours.
comment:4 by , 17 years ago
You can (temporarily) download the source SDF file that I am having this problem with here:
http://www.jasonbirch.com/temp/NanaimoCadastre.zip
It is pretty easy to reproduce, but specifically have a look at the Legal_Parcels feature type. GISLINK doesn't throw errors (numeric, possibly no nulls), but the next attribute down (legal_id maybe; not logged in at work right now) does.
comment:5 by , 17 years ago
Resolution: | → fixed |
---|---|
Status: | new → closed |
Well, I think it's fixed now, but you will have to test the theming -- I only did manual tests.
Was it an individual value theme or a bucket theme? In other words was it theming a number or a string? This will narrow it down to either select distinct or min/max computation.