Rowchart has Infinite height when _rowData.length is 0

[nordfjord]

See:

dc.js/src/row-chart.js

Line 169 in 6e95388

height = (_chart.effectiveHeight() - (n + 1) * _gap) / n;

when you filter such that a rowchart has no rows, its height becomes Infinity, which translates to a lot of errors along the lines of:

Error: Invalid value for <g> attribute transform="translate(NaN,0)"

[gordonwoodhull]

Seems like a legitimate bug, reopening (?)

[nordfjord]

It might be a legitimate bug.

But the errors I was receiving had nothing to do with this, I was way off base.

They were caused by a preRedraw function I had which calculated the xAxis domain and didn't check for infinity/NaN. I haven't seen any errors crop up around the height being Infinity, so I think it should be safe to close.

Regards,

Einar

[alisabzevari]

I think this is related to a calculation in updateElements function (https://github.com/dc-js/dc.js/blob/develop/src/row-chart.js#L173). This error happened to me:
Invalid negative value for <rect> attribute height="-1.3095238095238095"
It happened when I had too many rows and the chart has a small height. So the calculation made the height of every row to become negative.

[gordonwoodhull]

I looked and looked, but couldn't find any way to detect invalid DOM attributes being set in tests? Seems really weird that this can't be tested (via Jasmine). However, I did add a test and verified that it was creating errors before the fix (described in #1024) but not after.

If anyone knows how to test this, I'd really appreciate a hint!

[gordonwoodhull]

@alisabzevari, that's actually a different error - the opposite error. This one has to do with having no rows, not lots of rows.

If you'd care to file a separate issue, I'd appreciate it. Even better with a test case (fiddle or jasmine).

[gordonwoodhull]

Fixed in 2.0 beta 32