[indexPattern] Added #toDetailedIndexList() #5638
Conversation
| @@ -177,18 +177,21 @@ define(function (require) { | |||
| return this.intervalName && _.find(intervals, { name: this.intervalName }); | |||
| }; | |||
|
|
|||
| self.toIndexList = function (start, stop, sortDirection) { | |||
There was a problem hiding this comment.
Rather than adding complexity to this function by making it polymorphic, how about we move all of this logic into a new toDetailedIndexList function, and then modify the existing toIndexList to do something like:
self.toIndexList = function (start, stop, sortDirection) {
return this.toDetailedIndexList()
.then(indices => return _.pluck(indices, 'index'));
};
spalger
force-pushed
the
implement/indexPatternToIndexListDetailedResponse
branch
2 times, most recently
from
64ec792
to
6805797
Compare
spalger
changed the title
spalger
changed the title
|
Can you give a bit of a summary for the best way to test this? I can review the code and such, but I'm not really sure what behaviors I should be testing in kibana itself. |
|
Or is there a ticket associated with this that describes the issue? |
|
There isn't actually any issue, and the new method isn't used in Kibana at the time being. This is an API to support new functionality I'm working on. For now only the tests test it. |
spalger
force-pushed
the
implement/indexPatternToIndexListDetailedResponse
branch
2 times, most recently
from
df41329
to
2825c8d
Compare
spalger
force-pushed
the
implement/indexPatternToIndexListDetailedResponse
branch
from
2825c8d
to
e63838d
Compare
|
Ah, that makes a lot more sense. |
| .value(); | ||
| // TODO: remove when we get to es 2.2, see elastic/elasticsearch#14404 | ||
| let min = field.min_value; | ||
| if (typeof min === 'string') min = moment(min).valueOf(); |
There was a problem hiding this comment.
Will these definitely always be strings? Or more specifically, is it possible for them to be integer timestamps?
There was a problem hiding this comment.
After we upgrade to es 2.2, elasticsearch will always pass back unix timestamps. Until then times are passed back in whatever format they are stored in, and our only option is to attempt to parse that with moment()
There was a problem hiding this comment.
This makes complete sense. I totally had this reversed in my head.
|
If those constraints will always be strings or moment objects, then I'd say this LGTM. |
|
LGTM |
…stDetailedResponse [indexPattern] Added #toDetailedIndexList()
spalger
deleted the
implement/indexPatternToIndexListDetailedResponse
branch
indexPattern#toIndexList()now has adetailedResponseWe need access to the bounds for each index, so this new parameter returns the indices as objects rather than strings, and each object details the index, start, and end for each index.