Fix minor bugs in Perspective datetime + NaN handling #1028
Conversation
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
texodus
approved these changes
May 6, 2020
| def test_table_datetime_infer_from_string_with_time(self): | ||
| data = {"a": ["11:00 ABCD"]} | ||
| tbl = Table(data) | ||
| assert tbl.schema() == {"a": str} |
There was a problem hiding this comment.
Yeah, we definitely need to unify on one implementation of this and port it to C++
| # Augment the default error message when invalid values are | ||
| # detected, i.e. `NaN` | ||
| if error_string == "Out of range float values are not JSON compliant": | ||
| error_string = "Cannot serialize `NaN`, `Infinity` or `-Infinity` to JSON." |
There was a problem hiding this comment.
Very user friendly! Tragic that we must resort to this tho ...
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This PR fixes a few small bugs in
perspective-python:NaN,Infinity, and-Infinityare no longer serialized into JSON byPerspectiveManager, and any remote calls intoPerspectiveManagerthat returns a result withNaNinside will be rejected on the client-side. The server will not raise an Exception, but the exception will be visible inside the client browser console.A Short Explanation
Perspective's data loading and model treats
Noneornullas invalid values, and the Javascript valueundefinedas a value for "nothing exists in this cell". To the extent possible, Perspective will strip and treatNaNs asNone(in Python) ornull(in Javascript).A case where
NaNs cannot be coerced toNoneis in Arrows with columns that containNaNvalues but do not have thenull_countor the Arrow array's validity bitmap correctly set. To take advantage of the speed benefits provided by Arrow, Perspective treats Arrow-formatted data as strongly typed within each column, and where invalid values are indicated through use of the validity bitmap. This allows Perspective tomemcpyarrays from Arrow into Perspective, i.e. a basically free operation that massively speeds up Arrow loading.At this time, there are no plans to support a distinction between
NaNand None in Perspective, especially as the visualization value ofNaNandNoneare equivalent - both denote invalid values. Thus, usingperspective-pythonand attempting to serialize data containingNaNvalues will fail, and this change provides an explicit error message.Additionally, PyArrow's default behavior when converting from Pandas is to treat
NaNvalues asNone.Other Changes
Datetimes before the year 1900 are treated automatically as UTC using
calendar.timegm, instead oftime.mktimewhich raised an error whenever dealing with years before 1900.datetime.minvalues passed in to Perspective used to return "2001/1/1" due to a edge case in how timestamps were calculated before being passed into C++. Now, datetime.min values (whenever year, month, and day are1and hour, minute, seconds are0) return a timestamp of 0, i.e. the Unix epoch.