0.5.3 (#66)

• ⚡️ Optimize dplyr.arrange when data are series from the df itself
• 🐛 Fix sub-df order of apply for grouped df (#63)
• 📝 Update doc for argument by for join functions (#62)
• 🐛 Fix mean() with option na_rm=False does not work (#65)

0.5.2 (#61)

More of a maintenance release.

• 🔧 Add metadata for datasets
• 🔊 Send logs to stderr, instead of stdout
• 📌Pin dependency versions
• 🚨 Switch linter to flake8
• 📝 Update some docs to fit datar-cli

0.5.1 (#59)

• Add documentation about "blind" environment (#45, #54, #55)
• Change base.as_date() to return pandas datetime types instead python datetime types (#56)
• Add base.as_pd_date() to be an alias of pandas.to_datetime() (#56)
• Expose trimws to datar.all (#58)

0.5.0 (#52)

Added:

• Added forcats (#51 )
• Added base.is_ordered(), base.nlevels(), base.ordered(), base.rank(), base.order(), base.sort(), base.tabulate(), base.append(), base.prop_table() and base.proportions()
• Added gss_cat dataset

Fixed:

• Fixed an issue when Collection dealing with numpy.int_

Enhanced:

• Added base0_ argument for datar.get()
• Passed __calling_env to registered functions/verbs when used internally (this makes sure the library to be robust in different environments)

0.4.4 (#49)

• Adopt varname v0.8.0
• Add base.make_names() and base.make_unique()

0.4.3 (#46)

• Adopt pipda 0.4.5
• Make dataset names case-insensitive;
• Add datasets: ToothGrowth, economics, economics_long, faithful, faithfuld, luv_colours, midwest, mpg, msleep, presidential, seals, and txhousing
• Add base.complete_cases()
• Change datasets.all_datasets() to datasets.list_datasets()
• Make sure assume_all_piping mode works internally: close #45

0.4.2 (#44)

• Adopt pipda 0.4.4
• Add varname to dependency to close #30
• Rename datar.datar_versions to datar.get_versions
• Port a set of functions from r-base, incluing:
  • prod, sign, signif, trunc, exp, log, log2, log10, log1p,
  • is_finite, is_infinite, is_nan,
  • match,
  • startswith, endswith, strtoi, chartr, tolower, toupper,
  • max_col

0.4.1 (#43)

• Don't use piping syntax internally (>>=)
• Add python, numpy, executing and datar version to datar.datar_versions()
• Fix #40: anti_join/semi_join not working when by is column mapping

0.4.0 (#39)

• Adopt pipda v0.4.2

Performance improved:

• Refactor core.grouped to adopt pandas's groupby
• Try to use DataFrame.agg()/DataFrameGroupBy.agg() when function applied on a single columns (Related issues: #27, #33, #37)

Fixed:

• Fix when data or context as new column name for mutate()
• Fix SettingwithCopyWarning in pivot_longer
• Use regular calling internally to make sure it works in some cases that node cannot be detected (ie Gooey/%%timeit in jupyter)

Added:

• datar.datar_versions() to show versions of related packages for bug reporting.

0.3.2 (#34)

• Adopt pipda v0.4.1 to fix getattr() failure for operater-connected expressions (#38)
• Add str_dtype argument to as_character() to partially fix #36
• Update license in core._frame_format_patch (#28)