[pandas] annotated tag v0.8.0b1 created (now 130e52e)
Andreas Tille
tille at debian.org
Wed Dec 28 15:20:28 UTC 2016
This is an automated email from the git hooks/post-receive script.
tille pushed a change to annotated tag v0.8.0b1
in repository pandas.
at 130e52e (tag)
tagging 5b98fdfeae2ef42d78242c4cca1e0959b9c710db (commit)
replaces v0.7.3
tagged by Wes McKinney
on Mon May 28 21:06:01 2012 -0400
- Log -----------------------------------------------------------------
Version 0.8.0 beta 1
Adam Klein (157):
ENH: basic datetime64 integration and tests
changed vb_suite settings to my repo
Added datetime c extension skeleton
Got very basic skeleton of Date wrapper in cython up and compiling
Fixed boxing error
added some notes, cache dts in Date constructor
comments
make datetime64 array/asarray calls compatible with numpy 1.7
added datetime_helper to pull numpy 1.7 code we need
fixed gitignore
sketched fast accessor for grouping
ENH: added additional datetime support, backporting some numpy 1.7 code
additional parsing, accessors, Instant->Timestamp
small change to Duration constructor
ported dateutil relativedelta to cython, sketching idea for filter class
small fix to Delta
Ripped timeseries code from scikits.timeseries, got it to compile
fixed typo
cannibalizing scikits timeries
cannibalizing scikits timeries, missed a few files
added NP_LICENSE.txt
fixed module compilation issues
hacking on groupby
added field extractors, testing faster accessors
hacking on timestamp box
cleaned up a bit
working on box type craziness
more tests passing, mostly incompatible timestamp/datetime stuff
fix ordering in comparison
fix pyrex gdb flag
started work on freq integration in datetimeindex
Fixed Timestamp datetime wrapper
working on timeseries ranges
Added range calculations for monthly, daily, etc.
Work on frequency logic and integration with DatetimeIndex
added conformity check
Added relative monthly range, cleaned up here and there
datetime enhancements, cleaned offset logic, made datetimecache backed by khash for quick index lookup
changed offset design
Enhanced datetime to allow non-contiguous freq data
added cache extension logic
Fixed cache behavior with respect to start & end dates
cleaned up a bit, fixed anchoring logic in cache extension
added scikits.timeseries python code to codebase
Merge corrections
killed some bugs from merge
removed tests that snuck back in from merge
fixed shift conflict with old code
ENH: added SeriesBinGrouper, takes bin edges instead of labels
fixed quarterly offset fix, added small test
BUG: pyrex-gdb shouldn't be on by default
BUG: fixed possible bad freq access, added asobject method
BUG: fixed sorted set diff of DatetimeIndex
ENH: added group sum/mean/var by bins, with tests
ENH: started ripping out DateRange, replace with DatetimeIndex
BUG: Fixed merge to master
ENH: added offset attribute to DatetimeIndex
TST: added pickled timeseries and frame for testing
ENH: made DatetimeIndex pickle-able
ENH: import Timestamp directly
Merge branch 'master' into datetime64
BUG: typo
ENH: nuked DateRange, pulled into DatetimeIndex
BUG: removed stray import pdb
BUG: fixed +=/=+ error
BUG: fix pickling on Timestamp, all tests pass
ENH: added access to new offsets & deprecated old, unified offset/freq parameter
TST: remove object/dt64 test duplication, fix broken tests
ENH: pretty-print offsets where possible
BUG: fix offset names, relativedelta namespace issue
Merge branch 'master' into mergemaster
BUG: fix concatenation issue for M8 & object indices
BUG: fix extraneous comment
ENH: added simple fancy-indexing for datetime64 per #222
ENH: enable fancy datetime indexing for setting values, per #222
TST: added dt64 accessor tests, sketched additional accessors
ENH: added day-of-week, week-of-year, day-of-year, quarter accessors
ENH: went back to datetimestruct for field access, added sarray builder
ENH: first pass through custom grouper for datetime binning
ENH: cleaned up time/groupby API, expanded tests
ENH: fast offset generation in DTI for regularly strided offset
ENH: clean up unused counts variable
ENH: added numkeys to grouper object per Wes comments
ENH: added toOffset convenience, scetching freq convert method
ENH: implement basic convert function using groupby
TST: added a bit of testing around convert
ENH: convert handles up and down sampling
ENH: attach offset to converted timeseries, fix camel case issues
TST: fixed reference to legacy time rules
ENH: deprecate timeRule and offset in shift method, clean up cython build
ENH: deprecate offset/timerule for freq, add future warnings
ENH: fast time zone conversions and validation
BUG: fix scoping of exported variables, compilation unit conflicts on mac build
ENH: fancier datetime slicing logic, commented out b/c broke other tests, will fix
ENH: fancing time indexing working, cleaned variables per compiler warnings
ENH: merge conflict resolution
BUG: fixed failed tests from merge
ENH: merged up to lastest master
BUG: fixed merge problems from latest master
ENH: handle legacy unpickling for daterange-based objects
ENH: enable DatetimeIndex from string array
ENH: add vectorized datetimeindex construction from string array
ENH: a couple notes on IntervalIndex, TimeInterval
ENH: ripped all interval calculation logic from scikits.timeseries, continue sketchy interval object
ENH: added pretty interval printing
TST: made interval constructor test work
TST: fixed error raising up to python from skts code
BUG: fix cython compiling issue regarding error checking
BUG: fix behavior of Interval.now
ENH: added keyword constructors to Interval
ENH: added Ts convenience class, start sketching out IntervalIndex
ENH: fleshed out more of IntervalIndex
BUG: fix overflowin interval construction
ENH: add very basic interval slicing to intervalindex
ENH: add basic skts frequency multiples functionality, eg 5Min intervals
BUG: fix treatment of 'how' in asfreq having multiples
BUG: drop unused constant
ENH: merge up to master
ENH, TST: added interval accessors, port skts testing over
TST: added test_interval file
ENH: cleaned up interval accessor code
ENH: merged up to master
BUG: fix error from merging
ENH: asfreq -> resample, added index method, cleaned up tests
ENH: remove skts extraneous code from code base
ENH: added accessors to IntervalIndex
TST: added testing for intervalindex
BLD: fix setup.py clean not to remove skts.c
ENH: added interval/dti conversion code, tests
BUG: fix wrong to_offset call
ENH: added interval->dt64->interval conversions
ENH: delete accumulated datetime cruft
ENH: added strftime, working to adapt tsplot.py for IntervalIndex
ENH: added guards against bad interval data
ENH: got basic plotting on IntervalIndex'd series working
ENH: clean interval freq representation handling
BUG: fix tseries build on python 3 issue, closes #947
ENH: added dumb snap function to datetimeindex, to get to nearest offset
ENH: added test for datetimeindex snap
BUG: pass native char* instead of PyObject* to let cython handle 2/3 compatibility
BUG: remove PyString_FromFormat call, close #949
ENH: simplified unnecessary quarterly complexity, added additional aliases for intervals
BUG: fix freq access in tsplot
ENH: remove tsplot -> tsplotting, make work with intervalindex
ENH: git add tsplotting, fixed formatting of docstrings
ENH: clean up tsplotting.py
ENH: simple intervalindex plot check
TST: added basic interval plot test
ENH: check if datetimeindex can be plotted as intervalindex
ENH: add Q parsing, adding test to interval accessor to fix
BUG: fixed interval constructor when passed strings
ENH: merge up to master
ENH: added quarter-based slicing to datetimeindex
ENH: added basic DatetimeBlock to internals
TST: add note about what to test, clean extra code path
BUG: fixed YearEnd offset for month != 12, added tests
ENH: #1020 implementation. needs tests and adding to API
Chang She (174):
fixed 4Q1984 turning into 4Q3984
ENH: regularized frequency string, removes references to dt. in format, removes multiple if 1
allows for aliased 'how' argument in resample
ENH: Interval.resample allows lower case and start/end aliases. added time rule string to each DateOffset class as DateOffset.rule_code
ENH: to_timestamp supports 'S', 'E' end argument. aliased to start_time and end_time
TST: Interval freq string suppresses number if 1
ENH: key in HDFStore
ENH: call subplots_adjust on all DataFrame plot types to make it more friendly for figure reuse.
ENH: make empty histogram subplots invisible
BUG: _nanvar may return small negative if given a constant array
TST: typo in _check_ew_ndarray that skips int to float conversion check
ENH: add legend for subplots. GH #1108
ENH: allows Index - timedelta to make it symmetric to Index + timedelta
BUG: keep frequency information when adding timedeltas
Removed old duplicated makeDateIndex definition
ENH loffset in resample. GH #1127
Accidentally removed__sub__
ENH: pass single label to drop. GH #1143
TST: additional test case for tuple index entry
BUG: PeriodIndex tries to take frequency from first element. GH #1118
BUG: removed incorrect test case for GH #1118
BUG: default value for freq should be None
BUG: isnull(DatetimeIndex) raises exception
VB: TimeSeries.asof
ENH: TimeSeries.asof to take an array of dates/index
ENH: turn off redundant tick labels if sharex and sharey GH #1124
test script that cleans and builds extensions before running nosetests
Putting old Series.asof code back in just for scalar values
BUG: using 'E'.encode('ascii') instead of b'E' for Python25 compat
More Python 2.5 syntax compat
BUG: Frequency strings don't work with loffset parameter in resampling GH #1163
BUG: reimplemented timedelta.total_seconds for Python26 compat
TST: removed deprecated datetools import from test case
switch to equivalent str_to_bytes in py3compat
CMP: py3compat issues with StringIO. Imports should no longer fail in py3x, but writing non-unicode maybe a problem
ENH: small refactor of tseries.plotting and adding specialized PeriodIndex/DatetimeIndex plotting back in
added files with # in name to gitignore
BUG: tsplot calling Series constructor wrong when converting from DatetimeIndex to PeriodIndex.
ENH: column name option to Series.reset_index. GH #1164
BUG: negative numbers in freq str. GH #1162
ENH: Storing PeriodIndex in HDFStore. GH #1147
Merging minor plotting tweaks from GH #1056
TST: forgot to push the new testing module
BUG: fall back to regular plotting if no frequency information
Cleaning up tools.plotting
TST: skip timeseries plotting tests if no matplotlib
BUG: moved tseries.plotting import into test case to avoid early matplotlib import
TST: convenience funcs for period frame
ENH: shifting with PeriodIndex GH #1157
BUG: removed reference to unimported datetools in tseries/index.py. added test case for shifting with freq string
ENH: tshift method GH #1111
CMP: add __hash__ to make Period hashable in Py3
API: ordinal keyword in Period constructor. BUG: Period parsing monthly short string. GH #1170
BUG: overriding ptp in Series to work with np.ptp. GH #1166
TST: checks PeriodIndex field attributes against element-wise field attributes. GH #1146
ENH: first and nth for groupby. GH #994
BUG: return all indices if grouped by empty; fixes segfault. GH #1048
BUG: coerce index into DatetimeIndex if coercing from Series to TimeSeries
BUG: writing heterogenous Index to HDFStore. GH #1078
BUG: storing heterogenous Index types in HDFStore. GH #1078
DOC: one line example for GroupBy.size()
DOC: time_rule deprecation in sphinx docs for moving window functions #1042. Also started on new timeseries docs.
BUG: DataFrame.take raises exception on list input. GH #1196
DOC: documentation for Take methods. GH #1060
Add fields from DatetimeIndex to Timestamp. GH #1115
TST: fill with limit in sparse. BUG: limit in sparse frame. GH #1197 #993
ENH: Panel.transpose #695
ENH: copy keyword to Panel swapaxes/transpose #695
DOC: groupby.transform examples
ENH: sparse data structures in HDFStore. #85
BUG: unicode index handling in HDFStore. GH #626
ENH: limit keyword to Series/DataFrame align methods #216
VB: DataFrame multiaxis reindex. #836
DOC: more filtering examples. #746
ENH: as_multi keyword in Series/DataFrame constructor. #831
ENH: parsing numbers with commas in read_csv/table/clipboard/fwf #796
Revert "ENH: parsing numbers with commas in read_csv/table/clipboard/fwf #796"
Revert "ENH: as_multi keyword in Series/DataFrame constructor. #831"
ENH: parsing numbers with commas in read_csv/table/clipboard/fwf #796
TST: pass cmd line args to test scripts so can append -v etc
ENH: add comments keyword to text readers. #962
ENH: error message with malformed CSV file #1105
ENH: print err with row number with malformed CSV. #1105
BUG: alias offset freq to closest period freq when plotting #1202
DOC: DateRange -> date_range in Sphinx docs
BUG: unhandled case in Series with index=list of arrays #831
DOC: Sphinx docs creating Series with list of arrays as index to automatically create MultiIndex #831
ENH: to_timestamp and to_period for DataFrame #1191
ENH: cleaner implementation for #1191
DOC: rolling apply example #1210
BLD: platform int fixes
Stop storing class reference in HDFStore #1235
removed extraneous IntIndex instance test
ENH: convert multiple text file columns to a single date column #1186
Merged extra keyword with parse_dates
TST: VB for multiple date columns
A few related bug fixes
ENH: maybe upcast masked arrays passed to DataFrame constructor
VB: more convenience auto-updates
VB: get from and to email addresses from config file
VB: removing cruft; getting config from user folders
DOC: function for auto docs build
DOC: removed lingering sourceforge references
DOC: removed lingering timeRule keyword use
TST: test cases for replace method. #929
ENH: Series.replace #929
ENH: DataFrame.replace and cython replace. Only works for floats and ints. Need to generate datetime64 and object versions.
ENH: finishing up DataFrame.replace need to revisit
removed bottleneck calls from replace
moved mask_missing to common
TST: extra test case for Series.replace
removed remaining references to replace code generation
TST: rephrased .keys call for py3compat
Merge branch 'patch-1' of github.com:invisibleroads/pandas into pr1247
DOC: put back doc regarding inplace in rename in anticipation of feature
Merge github.com:invisibleroads/pandas into pr1248
DOC: reworded description for MultiIndex
DOC: started on timeseries.rst for 0.8
BUG: DataFrame.drop_duplicates with NA values
use fast zip with a placeholder value just for np.nan
TST: vbench for drop_duplicate with skipna set to False
optimized a little bit for speed
ENH: inplace option to DataFrame.drop_duplicates #805 with vbench
ENH: inplace keyword to DataFrame.rename #207
ENH: inplace Series.rename
ENH: inplace option for sort_index
TST: check cache clearing for inplace sort_index and rename
ENH: option to keep component date columns. #1252
ENH: squeeze keyword to return single col frame as series. #1253
ENH: several custom date conversion methods. #1174
DOC: started on docs for parse_dates and date_parser
DOC: documentation for parsing date/time from multiple columns #1250
ENH: custom date converters #1174
BUG: ts plotting fix after nanoseconds change #1265
BUG: align Series repr when unicode index is present #1279
ENH: intraday time filter #1275
TST: test length of filtered series
TST: tests for index column by name in read_table
ENH: allow index_col to be named in read_csv etc #1294
TST: using multiple date col as index
ENH: use parsed date column as index #1251
BUG: series/frame describe fails for datetime64 data
TST: test case for using frame column names in legend title f826c0e443
TST: using pandas.util.compat for py25 compat
ENH: pct change for Series and DataFrame #1271
ENH: flex comparison operators on DataFrame #652
TST: tests for flex compare #652
TST: minor change to test case for Series.describe
DOC: more notifications for daily doc build
DOC: fixed bug in auto_dev_build in docs make.py
BUG: raise exception when merging with insufficient info #1307
TST: equality of timestamps with diff time zones #1260
TST: additional test coverage for 0.8
TST: more tests for 0.8
TST: misc test coverage #1245
BUG: fixed take_2d_multi bug if dtype not in func dict
TST: adding to test coverage for core modules and minor fixes
BUG: minor bug fix for replace
TST: more release test coverage and minor fixes
TST: test coverage for parser
BUG: reading datetim64 dtype back in correctly
TST: additional test coverage for pytables and parsers
TST: non-unique monotonic index join
removing unused _get_indexer_standard
BUG: replace bug with non-float and non-object dtype blocks. Also bug in replace caused by np.isscalar(datetime) => False
TST: panel and internals test coverage
BUG: tsplot fix with business freq #1332
DOC: starting deeper revamp of ts docs for 0.8
DOC: more ts docs. Need to do resampling then PeriodIndex
DOC: rough draft of DatetimeIndex, date_range, shifting/resampling etc
DOC: timezone handling and started on Period
DOC: draft of timeseries section of docs. Added Period related documentation and examples
ENH: tz_convert for DataFrame #1330
DOC: groupby drop duplicate index #1312
Eric Chlebek (2):
BUG Deprecating use of bool() on DataFrame, GH #1069
Prevent tuple sublcass instances from being recognized as "list-like".
Kamil Kisiel (1):
BUG: Allow indexing with namedtuple, close #1026
Kelsey Jordahl (1):
ENH: Allow different number of rows & columns in a histogram plot
Kieran O'Mahony (5):
ENH: Add JSON export option for DataFrame take 2
ujson convert tabs -> spaces
ujson plug some memory leaks
ujson fix compiler warnings
ujson Python 3 support
Luca Beltrame (8):
ENH: Add support for converting DataFrames to R data.frames and
BUG: Properly handle the case of matrices
ENH: Add support for converting DataFrames to R data.frames and
BUG: Properly handle the case of matrices
Use NA_Real, or conversion to matrices will turn NAs into zeros
BUG: Use NA types for better conversion, and refuse matrix conversions
DOC: Add documentation for conversion to R objects
DOC: Forgot a rather important "no"
Mark Wiebe (5):
Remove dependencies on details of experimental numpy datetime64 ABI
Use datetime64 with a 'us' unit explicitly, for 1.6 and 1.7 compatibility
Use an explicit unit for the 1.7 datetime64 scalar constructor
Use assert_equal instead of assert, to see the actual values
Microseconds (us) not milliseconds (ms)
Peng Yu (1):
fixing the error in io/sql.py
Roy Hyunjin Han (2):
Removed erroneous reference to iterating over a Series, which iterates over values and not keys
Fixed a few typos
RuiDC (2):
treat XLRD.XL_CELL_ERROR as NaN
replace tabs with spaces
Skipper Seabold (13):
ENH: Add explicit init with month check to YearEnd
ENH: Add new offsetMap
TST: Add tests for BYearBegin offset. Make assertEq more verbose on failure.
TST: Remove test for quarter begin in month 1-3
ENH: Add BYearBegin offset. Fix new offset map. Remove month restriction on quarter offsets
TST: Add tests for MonthBegin
ENH: Add MonthBegin offset. Start QuarterBegin. Finish adding all time rule aliases
TST: Add tests for QuarterBegin
ENH: Add QuarterBegin Offset
TST: Add tests for BMonthBeing and BQuarterBegin
ENH: Add BMonthBegin and BQuarterBegin
STY: Make some comments and cleanup some logic
ENH: Add WOM to new offsets
Stefan van der Walt (1):
BUG: work around missing NPY_FR_B in numpy 1.7
Takafumi Arakaki (7):
ENH: treat complex number in maybe_convert_objects
ENH: treat complex number in maybe_convert_objects
TST: Add complex number in test_constructor_scalar_inference
ENH: treat complex number in internals.form_blocks
ENH: add internals.ComplexBlock
BUG: fix max recursion error in test_reindex_items
BUG: replace complex64 with complex128
Thomas Kluyver (8):
PyInt aliases for Python 3.
Alias basestring for Python 3 in Cython file.
Make Tick and subclasses hashable in Python 3.
Pickle should always read files as binary.
Fixes for passing characters to skts_resample in Python 3
Integers extracted from arrays can also be numpy scalars.
Use next() builtin rather than method.
Alias next() for Python 2.5
Vytautas Jancauskas (5):
Changes to plotting scatter matrix diagonals
Changed xtick, ytick labels
Added simple test cases
Updated plotting.py scatter_matrix docstring to describe all the parameters
Added scatter_matrix examples to visualization.rst
Wes McKinney (324):
ENH: start convenience date range functions
ENH: Working on getting to merge ready, cleanup. See full commit message
ENH: cython refactor, pushing more index logic into engines.pyx, broken unit tests, a mess still
ENH: index tests pass
BUG: make it further in the test suite
TST: more test coverage progress
BUG: temporary hack around TypeError in PyObjectHashTable
ENH: non-unique index support. much more testing needed
ENH: implement limit option to reindex/fillna methods and related infrastructure/API methods, doc updates still needed #825
DOC: docs for limit param, #825
fix merge conflicts
BUG: fix up borked merge conflicts. test suite passes now
ENH: preserve DatetimeIndex frequency in boolean arrays with single contigous chunk, close #791
ENH: can group by level 0, enables grouping non-unique indexes, close #643
TST: unit test for #466
TST: moved test_datetime64.py tests to test_timeseries
TST: remove errant foo and test_datetime64.py
TST: interactions between array of datetime objects and DatetimeIndex, bug fixes
ENH: enable to_datetime to be vectorized, handle NAs, close #858
ENH: use cython bin groupers, fix bug in DatetimeIndex.__getitem causing slowness, some timeseries vbenches
ENH: implement Cython OHLC function for groupby #152
ENH: integrate cython ohlc in groupby and test, close #152
BUG: convert datetime64 -> datetime.datetime for matplotlib, close #1003
ENH: initial version of convert_to_annual for pandas, #736
BUG: name inline method something different
BUG: fix broken time_rule usage in legacy DateRange, close #1036
ENH: handle datetime64 in block formation from dict of arrays in DataFrame constructor, close #1037
ENH: more datetime64 integration in core data algorithms per #996, close #1035
ENH: to_datetime will convert array of strings and NAs to datetime64 with NaT, close #999
ENH: legacy time rule support and refactoring, better alias handling. misc tests, #1041
ENH: add to_datetime method to Index, close #208
BUG: fix major performance issue in DatetimeIndex.union affecting join performance on irregular indexes, remedying #1046
Merge changes from master
RLS: set released to false, bump dev version to 0.8.0
BUG: fix buglets surfacing from merge
TST: don't use deprecated DateRange
Merge pull request #1038 from neurodebian/master
Merge branch 'timeseries'
BUG: enable numpy reductions to be passed to how parameter in convert, TimeGrouper tests, close #1045
ENH: more legacy time rule support, close #1041
ENH: partial date slicing with DatetimeIndex, close #1044
ENH/API: add normalize option to DatetimeIndex, date_range, bdate_range. deprecate /remove _normalizeFirst option in offsets. set default for BDay to not normalize, close #1031, #506
ENH: implement group_prod, group_prod_bin cython aggregators, close #1018
ENH: group_max/min and bin versions, close #1019
ENH: more TimeGrouper issues in apply, close #1057
ENH: can pass list of columns to parse_dates, close #853 and add dayfirst argument for european dates close #854
BUG: fixes to bugs raised by moving to Cython 0.16rc2
BUG: fix searchsorted off by one error, close #1068
ENH: can pass valid pytz timezone name to date_range, DatetimeIndex close #1089
BUG: treat empty string as NaT in to_datetime, close #1085
ENH: week of month legacy naming / cleanup close #1088
TST: failing unit test for #1063
ENH: group_position prototype function
BUG: incorrect comparison of Tick objects, close #1063
REF: major time series refactor, rename resample to asfreq, skts -> period, various cleanup
REF: remove empty finance subpackage
ENH: ship parts of scipy.stats, close #1092
BUG: better error message with misspecified date_range, fix buglet. close #1095
ENH: better KeyError when timestamp not found in time series, close #1062
ENH: make Day a 'tick' so it's faster
DOC: release notes
TST: test reindexing timeseries with datetime.datetime sequence, close #1099
ENH: add order method to Index, close #1028
ENH: Timestamp can accept date objects, promote datetime.date in Index.join/get_indexer, close #867
BUG: don't lose DatetimeIndex metadata when calling take, close #1101
BLD: fix circular imports, close #1104
BLD: add tseries/compat subpackages, data files to setup.py
BUG: raise exception with partial indexing on unordered time series, close #1080
REF: split time series resampling into resample.py, other miscellaneous refactoring
ENH: refactoring toward #1064, but broke something
ENH: refactor / cleanup to do timestamp -> period resampling, re #1064 #1004
BUG: fix out-of-bounds issues / bugs in groupby binners, close #1107
BUG: testing re #1004, fix bug in resample
ENH: implement first/last subset functions, more tests probably needed. close #865
TST: add offset to DatetimeIndex, close #1096
BUG: squeezing aggregated cython result incorrect logic when ngroups == 1, close #1065
TST: add unit test for downresample non-unique time series, per #1077
BUG: raise exception when trying to reindex non-unique index, test asfreq on non-unique timestamps close #1077
BUG: fix formatting of pre-1900 dates in Index.format, close #1100
REF: move tsplotting mess to pandas/tseries
ENH: add to_timestamp method for PeriodIndex->DatetimeIndex, fix skts bug, don't normalize time away in annual offsets
BUG: weekofyear/dayofyear performance problems #1112
BLD: use npy_int64 instead of system long, #1123
ENH: add frequency to Series repr and DataFrame.info, close #1102
ENH: parse mixed time strings like '1h30min15s', close #1120
BUG: override DatetimeIndex.repeat to drop frequency, close #1110
REF: remove code dup from field accessor function creation
TST: unit test for frequency inference function
ENH: more frequency inference, unique_deltas function
ENH: frequency inference and tests, close #391
BUG: implement take on PeriodIndex, close #1117
BLD: debugging windows debug problems
BLD: int types MSC compat
BLD: msvc fixes
BLD: 32-bit / MSVC fixes
BLD: convert indexer to int32
Merge takluyver/tseries-py3
BLD: python 3 fixes
ENH: make Period.to_timestamp more flexible, match PeriodIndex.to_timestamp, close #1116
BUG: test/fix to not always normalize with quarterly time rules, GH #1113
ENH: conversion to UTC between differently indexed datetime indexes, close #864
BUG: more tz conversion testing
ENH: proper types, speed up tz_localize
ENH: add lite frequency inference to DatetimeIndex setops, close #998
ENH: implement asfreq with PeriodIndex on Series/DataFrame, close #1129
Merge pull request #1140 from changhiskhan/loffset
ENH: frequency inference when adding delta to DatetimeIndex
TST: better error message, close #1118
ENH: can pass datetime.time to get values at time, close #1130
BUG: loss of freq in DatetimeIndex constructor, close #1140
ENH: joins, setops, alignment for PeriodIndex and objects having one as index, close #1138
ENH: add DatetimeIndex.normalize and is_normalized property, close #1132
ENH: basic downsampling with PeriodIndex, GH #1109
ENH: add to_period method to TimeSeries to convert DatetimeIndex -> PeriodIndex
ENH: start on is_subperiod function, still more for #1109
REF: refactoring timegrouper for upsampling
BUG: fix bugs in tests, more resampling
ENH: test cases for period upsampling, implement kind='timestamp' close #1109
BUG: use own period frequency when converting to timestamp
Python 3 compatibility in datetime string parsing, hacks around unorderable type errors from index engine
BLD: handle ascii encoding in cython, fix 32-bit segfault hopefully
BLD: get rid of parse_python_string for python 2.5
BUG: must cast to uint8 for python 2.5
BUG: cast mask back to bool (hack)
BUG; cast to uint8 for row_bool_subset
BUG: cast bool to uint8 in engines.pyx, skip namedtuple test in python 2.5
BUG: another bool uint8 cast, add period shortaliases after making reverse_period_code_map
REF: slightly more compact DatetimeIndex repr
API: add offsets to tseries.api
ENH: resampling annual to quarterly if end in same month
BUG: fix quarterly date to period conversions in Period, PeriodIndex
BUG: fix annual by-month period conversion
BUG: don't autofmt_xdate if use special ts formatter
BUG: pass right plot function to tsplot when logy=True, close #1193
REF: time_rule -> freq addition and API deprecation in moving window functions close #1042
ENH: speed up Series.__getitem__ with timestamps, much more efficient implementation of Series.asof, close #1168
REF: working on UTC only timestamp handling, timezone handling modifications, not yet working
ENH: convert UTC -> local, attach right tzinfo #1156
TST: check right tzinfo attached, close #1156
ENH: convert passed dates to UTC, add tz_convert method to Timestamp #1172
TST: test resample functions as fill_missing values, #1114
ENH: make DatetimeIndex + 1 do shift like PeriodIndex, close #1158
TST: vbench for #1152, some work on improving performance of asfreq
ENH: better vectorize asfreq conversion re: #1152
ENH: small amount of refactoring
ENH: finished fixing resampling period-indexed data, only 50% slower than timestamps now, close #1152
BUG: Series boolean cast bug and fix binary ops between boolean Series with NAs, close #1074
ENH: allow duplicates in index when parsing file, close #1173
TST: add tests for indexing Series with duplicate periods, close #1154
TST: only convert to list when non-list in safe_fetch per #1070
DOC: explicit check for NumPy >= 1.6.1 in setup.py, close #1061
DOC: improve docstring and add rst doc for #695
BUG: allow assigning non-scalars in mixed-type DataFrame, no index checks performed, close #1142
BUG: label slicing bugs with floats in Series, DataFrame, close #1167
REF: remove Picklable class
TST: raise ValueError when no group keys passed, close #1048
TST: unit test and close #522
BUG: enable .ix to better handle indexes with duplicates, close #1201
Merged PR #1073
RLS: note about #1073
BUG: imprecise logic leading to malformed Series in Series.apply, close #1183
ENH: add Series/DataFrame.update methods and rst docs, close #961
ENH: pass list of array to create MultiIndex, close #831
ENH: can pass external arrays in lieu of column names to DataFrame.set_index, close #402
ENH: groupby refactoring to enable reasonably fast resampling of panel data, unit test for user-specified function fails #1149
ENH: hack job but panel resampling with NumPy function works, close #1149
BUG: support PeriodIndex as DataFrame columns and deal with other PeriodIndex support issues in Series/DataFrame constructors, close #1211
ENH: unstack multiple columns in one shot to eliminate empty columns in pivot table operations, close #1181
ENH: unstack multiple levels in one shot per #1181
BUG: raise TypeError when comparing None with float series, unit test to close #1079
TST: verify Index of datetime.datetime -> DatetimeIndex, close #1187
ENH: special case reindexing both axes on homogeneous DataFrame for major speedup, vbenchmarks #836
ENH: special case square reindexing of DataFrame in .ix. not pretty but close #836
BUG: fix multiindex formatting problem close #1217
TST: fix red herring test failure on win32 by passing int explicitly #1219
TST: fix 32-bit/python < 2.7 test failures
REF: great int32 -> int64 migration, though 0.01% chance of success on 32-bit
BLD: more platform int fixes per #855
BLD: 32-bit take failures
TST: more take int32 compat
ENH: always sort pivot table columns per #1181
BLD: time.h hack for mscvr90.dll in mingw
BLD: remove time.h from np_datetime.c
BLD: store labels as platform int in MultiIndex
BLD: more int32->int64 on 32-bit fixes
ENH: handle non-DatetimeIndex to support statsmodels
TST: add generated.c to count_code.sh
REF: working toward #1150, broke apart Cython module into generated _algos extension
REF: have got things mostly working for #1150
BUG: more bug fixes, have to fix intraday frequencies still
BUG: more intraday unit fixes
BUG: test suite passes, though negative ordinals broken
BUG: weekly and business daily unit support #1150
REF: remove period multipliers, close #1199
ENH: move _ensure_{dtype} functions to Cython for speedup, close #1221
DOC: doc fixes
ENH: handle dict return values and vbench, close #823
ENH: add is_full method to PeriodIndex close #1114
TST: use NaT value
ENH: add docs and add match function to API, close #502
ENH: add Cython nth/last functions, vbenchmarks. close #1043
BUG: fix improper quarter parsing for frequencies other than Q-DEC, close #1228
BUG: implement Series.repeat to get expected results, close #1229
ENH: anchor resampling frequencies like 5minute that evenly subdivide one day in resampling to always get regular intervals. a bit more testing needed, but close #1165
BUG: support resampling of period data to, e.g. 5minute thoguh with timestamped result, close #1231
BUG: remove restriction in lib.Reducer that index by object dtype. close #1214
TST: vbenchmark for #561, push more work til 0.9
BUG: don't print exception in reducer
BUG: rogue foo
ENH: reimplment groupby_indices using better algorithmic tricks, associated vbenchmark. close #609
BLD: fix npy_* -> pandas_*, compiler warnings
TST: remove one skip test
ENH: store pytz time zones as zone strings in HDFStore, close #1232
BUG: fix rebase conflict from #1236
RLS: release note
TST: test with headers
RLS: release notes
ENH: optimize join/merge on integer keys, close #682
RLS: release notes for #1081
ENH: efficiently box datetime64 -> Timestamp inside Series.__getitem__. close #1058
BLD: add modified numpy Cython header
BLD: fix datetime.pxd
ENH: can pass multiple columns to GroupBy.__getitem__, close #383
ENH: accept list of tuples, preserving function order in SeriesGroupBy.aggregate
ENH: more flexible multiple function application in DataFrameGroupBy, close #642
DOC: release notes
BLD: fix platform int issues
TST: verify consistently set group name, close #184
ENH: don't populate hash table in index engine if > 1e6 elements, to save memory and speed. close #1160
ENH: support different 'bases' when resampling regular intervals like 5 minute, close #1119
BUG: floor division for Python 3
ENH: very basic ordered_merge with forward filling, not with multiple groups yet
ENH: add group-wise merge capability to ordered_merge, unit tests, close #813
BUG: ensure_platform_int actually makes lots of copies
RLS: release notes, close #1239
BLD: 32-bit compat fixes per #1242
ENH: add keys() method to DataFrame, close #1240
DOC: release notes
DOC: release note re: #929
REF: microsecond -> nanosecond migration, most of the way there #1238
BUG: more nano fixes
REF: more nanosecond support fixes, test suite passes #1238
ENH: more nanosecond support #1238
DOC: release notes
ENH: add KDE plot from #1059
BUG: fix logical error in values_at_time, close #1249
TST: more careful check for malformed BlockManager, close #815
DOC: removed reference to unique Index. more docs work will be needed. close #1208 for now
DOC: handle pure python code blocks more correctly in ipython_directive, close #560
BUG: handle ndarrays/Index objects also in GroupBy.__getitem__, close #1257
RLS: release note for #1243
BUG: fix resampling buglet and weekly period ordinals, close #1259
BUG: Timestamp hashing, misc test coverage
TST: more test coverage, bug fixes. don't reference datetime_ for NumPy 1.7
BUG: nanoseconds for NumPy 1.7
BUG: don't import scipy at import time, close #1262
BUG: more NumPy 1.7 fixes
BUG/TST: more test coverage, resample bugfix
DOC: release note re: #207
BUG: can pass list of functions to SeriesGroupBy.aggregate, fixes #1268
ENH: more helpful error message when nothing passed to reindex, close #1267
ENH: add nanosecond field accessor to DatetimeIndex, close #1261
ENH: use DataFrame columns name for legend title in plotting code, add marker option to scatter_matrix
BUG: unicode repr handling fix for Python 3 re #1292, #1279
DOC: doc fix since sphinx doesn't like file://
DOC: fix typo in docs
BUG: fix CRLF and kludge around datetime64 nanosecond JSON encoding issues per #1263
ENH: return DatetimeIndex from to_datetime, control formatting of datetime64 values in DatetimeIndex.__repr__
ENH: add max_colwidth print option #1288
TST: fillna inplace only after 4/2012
TST: vbench always use fresh clone due to weird issues
TST: tuple -> datetime
TST: more vbench modules, some more start_date fixes
TST: typo
BLD: msvc fix (hopefully) + start resample benchmarks later in 2012
BLD: more msvc fixes
BLD: MSVC fixes, passes on 64-bit
BLD: win32 fix
TST: unit test for #1307
ENH: work toward joins on non-unique indexes, many-to-one support per #1306
ENH: implement many-to-many index joins, close #1306
DOC: release notes
BUG: handle other datetime64 scalar frequency units in Timestamp/DatetimeIndex, close #1314
BUG: fix infinite loop on win32/64
TST: start a couple vbenchmarks later
BUG: don't lose Index name when grouping by level=0, close #1313
ENH: get rid of dummy names in groupby indexes, close #1291
BLD: fix numpy 1.6/1.7 datetime API problems
ENH: ujson python 3 defines
BLD: add py_defines.h
BUG: fix several many-to-one join bugs
ENH: pct_change skip over NAs, close #1271
TST: add unit tests to verify close #1316
BUG: handle list of integers in DatetimeIndex constructor, close #1303
BUG: unpickling length-0 DateRange as Series.index, close #1274
BUG: always store UTC datetime64 value in Timestamp, test comparisons. close #1260
ENH: encode datetime objects as nanosecond timestamps using numpy machinery, close #1304
BUG: copy offset to DatetimeIndex in constructor
BUG: fix multitake special case with boolean arrays in .ix, close #1321
BUG: cast indexers to platform int for 32-bit systems
BUG: python 2.5 fixes
TST: skip failing UJSON tests on Python 2.5
ENH: create datetime arrays in type inference, more testing needed
ENH: conversion to datetime64, per #1317
TST: suppress Series warnings, speed up Series.__getitem__ in typical use cases
ENH: NaT type and integration throughout, NA handling. #1286
REF: rename pandas._tseries to pandas.lib, close #1315
BLD: add datetime_helper.h
ENH: review pandas.tseries.api, close #1169
TST: python 3 / json fixes re: PR #1326
ENH: return -1 in NaT field access, vector and scalar. close #1310
BUG: intercept NumPy 1.6 M8[ns] -> malformed O in DataFrame.values, Series.astype, etc. close #1323
ENH: mix arrays and scalars in DataFrame constructor, close #1329
BUG: alias W to W-SUN, add test for buglet close #1327
BUG: DatetimeIndex partial slicing bug, tsplot kludge around #1332
ENH: construct PeriodIndex from arrays of fields, allow negative ordinals. close #1333 and #1264
BUG: raise exception for unintellible frequency strings, close #1328
BUG: display localtime in DatetimeIndex.__repr__, close #1336
ENH: add NA handling to scatter_matrix, close #1297
TST: testing to close #1331
ENH: add option to use Series.values to interpolate, close #1206
RLS: release notes
RLS: 0.8.0beta
RLS: 0.8.0 beta 1
Wouter Overmeire (2):
BUG: fix to_html with colums MultiIndex without level names, close #1053
TST: make TestLegacySupport setUpClass independent of current path.
Yaroslav Halchenko (3):
ENH: @network decorator for tests
TST: split test_url and make with @network portion requiring network access
ENH: skip test_file test with python 2.5 (not supported AFAIK)
timmie (1):
replace statsmodels by pandas
-----------------------------------------------------------------------
No new revisions were added by this update.
--
Alioth's /usr/local/bin/git-commit-notice on /srv/git.debian.org/git/debian-science/packages/pandas.git
More information about the debian-science-commits
mailing list