Created
May 9, 2014 16:53
-
-
Save wboykinm/d22689cb5cade5baf733 to your computer and use it in GitHub Desktop.
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
[12:13] <wboykinm> strk: So I've got a table with a poorly-behaving column | |
[12:14] <strk> "poorly-behaving" ? how so ? | |
[12:14] <wboykinm> I'm not precisely sure what's wrong with it, but here's how I narrowed it down: | |
[12:14] <wboykinm> 1.) downloaded table as .shp | |
[12:15] <wboykinm> 2.) re-uploaded table as .zip (exactly as downloaded) | |
[12:15] <wboykinm> 3.) whole swaths of features are missing | |
[12:15] <strk> uhm | |
[12:15] <wboykinm> like this: https://www.flickr.com/photos/vtcraghead/14123997052/ | |
[12:15] <wboykinm> But wait, it gets better | |
[12:15] <wboykinm> So I download as .shp again, | |
[12:16] <strk> select count(*) gives the correct number ? | |
[12:16] <wboykinm> then I remove all columns except for my unique key, a string column | |
[12:16] <wboykinm> compress and re-upload, and it works perfectly - all features visible | |
[12:16] <wboykinm> last diagnostic: | |
[12:17] <wboykinm> I download as .shp, remove all columns except for the key column and the name column, and the missing-feature problem returns. | |
[12:17] <wboykinm> . . . leading me to believe the string "name" column is to blame. | |
[12:17] <wboykinm> does that make sense? | |
[12:18] <strk> I'm not very knowledgeable of the import process so can't tell exactly what's going on, but your debugging seems perfect | |
[12:18] <strk> how about producing a syntetic small version of the case and attaching to a github ticket ? | |
[12:18] <strk> would seem easy to produce | |
[12:18] <strk> hopefully the number of features doesn't affect it | |
[12:19] <wboykinm> will do. WHich repo should I attach it to? | |
[12:19] <strk> you're basically saying that the presence of a "name" column breaks the import, right ? | |
[12:19] <wboykinm> right | |
[12:19] <strk> github.com/cartodb/cartodb | |
[12:19] <wboykinm> but the encoding seems fine | |
[12:19] <wboykinm> No characters that cartodb hasn't handled before | |
[12:19] <strk> and "breaks the import" means "drops some rows", you said (but I'd confirm with SELECT count(*)) | |
[12:20] <wboykinm> okay, doing that now . . . | |
[12:20] <strk> it's important to distinguish between data and map issues | |
[12:22] <wboykinm> Yup. rows dropped. Should be 35603, only importing 33004 | |
[12:22] <strk> a-ha! | |
[12:22] <strk> see if you can produce a much smaller version of the test, like ideally 2 rows (of which 1 only imports) | |
[12:23] <strk> then the bug would probably become obvious ;) | |
[12:23] <wboykinm> okay, I'll target those... | |
[12:24] <strk> or failing to reduce in number of rows, reduce complexity in values, like make all "name" values the same, and see if it still occurs... | |
[12:24] <strk> I'm off for today, feel free to report on the googlegroup :) |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment