GEMS1A many participant observation anthropometry variables (all integer, categorical?) give 500 in Histogram and Boxplot

[bobular]

No filters, qa.clinepidb.org back end.

I tested all the anthropometry variables in the Histogram (main var) and Boxplot (y-axis, using country on x-axis), and it seems that all those measures involving weight and MUAC are giving 500 errors.

This may be related to variables derived from multiple valued variables?

I won't diagnose further as this is probably a known issue.

[danicahelb]

GEMS1A is loaded with updated code, I just checked and am seeing the same thing as bob. Also, "weight" is greyed out & cannot be plotted in either the boxplot (y-axis var) or histogram (main var)

[SheenaTomko]

This is because those variables are loaded as "integers" and there is a bug with integers not being accessible in the visualization tools. I reported that to Danielle here in this ticket that I think was supposed to allow them.

There's a separate issue that GEMS variables like Weight and MUAC were over-truncated when loading. They are both supposed to have one decimal place. That has been reported to Jay in the redmine tickets for loading studies.

@danicahelb and @bobular I would recommend closing this ticket

[danicahelb]

I'm going to leave this ticket open and add a paused label. We will retest once Jay has fixed the over-truncating issue

[bobular]

No I don't think we should close it. It seems integer categorical variables are still causing histogram to return 500. I'll change the title of this ticket

[steve-fischer-200]

@bobular i cannot reproduce this. please try and let me know.

[d-callan]

gems1a seems to be unavailable today. i looked at gems1 and can confirm that there at least integers work in the histo and box plot. there was a known issue w integers not working a week or so ago which has since been resolved, which is about the right time for this to have been reported. i also suspect we can close this.

[ryanrdoherty]

@bobular Can you test now? Gems 1A is up. As you suggested, there is a known issue with multi-value vars; if we find that this is working except parts that are a duplicate of https://github.com/VEuPathDB/EdaLoadingIssues/issues/12, we can close this one.

[bobular]

@ryanrdoherty GEMS1A is not there right now (using dfalke-b as back end) - could you Slack me when it's available for testing?

[ryanrdoherty]

@bobular Looks like GEMS1A is now available on qa.clinepidb.org

[bobular]

Those vars work fine now - yay!