Closed bcodell closed 1 year ago
@tnightengale looking for the following feedback:
feature_json
key as an included column a bug or did this functionality intentionally not exist? It seems like you intended to implement it since you had the json_unpack_key
helper function already in place (asking for semantic versioning purposes)feature_json
?example__activity_stream
(removing feature_json
from array, snake-casing json keys) reasonable?@tnightengale looking for the following feedback:
- Was not being able to specify a
feature_json
key as an included column a bug or did this functionality intentionally not exist? It seems like you intended to implement it since you had thejson_unpack_key
helper function already in place (asking for semantic versioning purposes)- Does the abstraction I put in place make sense, or are there changes you'd like to see?
- Does this assumption (baked into the code make sense: that if a specified included column isn't identified as one of the standard columns (or project-specific aliases) from the Activity Schema spec, it is assumed to be contained in the
feature_json
?- Are the semantic changes I made to
example__activity_stream
(removingfeature_json
from array, snake-casing json keys) reasonable?
parse_column
function? Yes I think it makes sense.Thanks @tnightengale! I'm going to merge this then open a PR to bump the version to 0.4.0. This feature should provide a workaround to #26 and #33 until those bugs are explicitly resolved.
This PR:
parse_column
which returnstable_alias.column
and extracts the column from the activity stream'sfeature_json
column if the column name is not identified as a standard column from the Activity Schema, then applies that function to all columns selected from all primary and appended columns in thedataset
macrofirst_ever_1
) to extract json features from the primary and appended activityexample__activity_stream
and relevantoutput__<dataset>
csv files so thatfeature_json
values aren't nested in an array and so thatfeature_json
keys use snake case formattingincluded_columns
argument in the Readme