--feat_group_by_outcomes

Switch

--feat_group_by_outcomes

Description

Creates a new feature table, one per distinct value of the given outcome(s), splitting the original feature table's rows by outcome value.

Argument and Default Value

No argument. Boolean switch, default: False.

Details

Known issue — broken in SQLite mode. FeatureRefiner.createFeatTableByDistinctOutcomes() calls mm.execute(...) directly with MySQL-flavored SQL (including a LIKE-based table-existence check). Verified: fails with sqlite3.OperationalError: near "like": syntax error. Works against MySQL.

Other Switches

Required Switches:

• -d, -g, -t

• -f

• --outcome_table, --outcomes

Optional Switches:

• --output_name (used as a suffix on the generated table names)

Example Commands

Requires MySQL — broken against SQLite, see Known issue above:

dlatkInterface.py -d dla_tutorial -t msgs -c user_id -f 'feat$1gram$msgs$user_id' \
--outcome_table blog_outcomes --outcomes is_student --feat_group_by_outcomes