Daft 0.0.14 Release Notes#

The Daft 0.0.14 release features many stabilization and bugfixes. The highlights are:

  • Adds index syntax for accessing columns (df["mycol"])

  • Null testing and hardening for numerical expressions

  • Null testing and hardening for aggregations/groupby operations

New Features#

DataFrame Indexing Syntax#

DataFrame columns can now be accessed via an indexing syntax, greatly improving code readability and ease of use.

Example:

df = df.with_column("foo_plus_1", df["foo"] + 1)
df = df.join(df2, left_on=df["id"], right_on=df2["id"])

See: #265

Global Aggregations#

DataFrame.agg added for aggregations on a non-grouped DataFrame

See: #263

Enhancements#

  • implement distinct logical plan and remove hack in aggregate #262

Bug Fixes#

  • Fixes and unit tests for aggregations on null items #268

  • Fix explode column pruning optimizer bug #257

  • Add null testing for operators and fix mod operator #247

  • Filter out nulls when running sample function for sort shuffle op #267

Build Changes#

  • Test Wheels after building to verify they work #260

Closed Issues#

  • Use indexing syntax for getting/setting columns in a Dataframe (e.g. df[“foo”]) #259

  • Explode column pruning optimizer bug #258

  • Run workflow that installs wheels to verify they work #253

  • Build Distinct Logical Plan Operator #245

  • Allow for .agg aggregations over a non-grouped DataFrame #243