NewYour coding agent can read the release notes before it upgrades.Set up the MCP server →
PyPI · #2394 most downloaded on PyPI
ODPS Python SDK and data analysis framework
Last release 7 days ago
10 Sep 2026
Ships fairly regularly
a new release about every 2 months
Nearly every release is documented
notes for 55 of the last 60 stable releases
1 version withdrawn
withdrawn after publishing
11 years old
142 releases · first in 2015
Support signing with bearer token.
- Add sql cost task support - Bug fixes
Support pass timezone as settings when submitting SQL tasks
One column per quarter.
RandomScalarrunning_clusterNothing published for this version
Support referencing resources in other projects when creating functions
Add support for new window functions (cumstd, qcut, nth_value, cume_dist)
quantile(x) in pivot_tableAdd fast tunnel for simple types (bigint, double, boolean) for pandas
Make PyODPS compatible with Python 3.7, changes including support for the new method invocation instruction as well as changing async args into async_
async args into async_ to avoid conflicts with the new keywordproject arg on odps.df.funcAdd support for percentile in DataFrame
Update cloudpickle. Now supports Python 3.6 and fixes WorkerCrashfile under several scenarios
Add support for lateral views (see this link for more details)
Add method filter_parts to DataFrame, to provide a more straightforward partition representation compatible with MaxCompute representation
Nothing published for this version
Add support for new types in MaxCompute 2.0 in basic API, including create table, upload/download data etc. UDF not supported
Support async and delayed execution of to_pandas()
DataFrame.query API, filter data by string, e.g. df.query('sepallength < 5')
df.query('sepallength < 5')Support nunique on multiple columns, e.g. df['a', 'b'].nunique()
df['a', 'b'].nunique()DataFrame support partial update on column, e.g. df[df.field < 10, 'col'] = 1
df[df.field < 10, 'col'] = 1Nothing published for this version
* Wheel support for installation * Bug fixes
Add options.tunnel.string_as_binary option to treat the string field as bytes
options.tunnel.string_as_binary option to treat the string field as bytesNothing published for this version
* Optimize tunnel performance * Bug fixes
Add tunnel optimization for filter_partition operation
Drop, setdiff, intersection, applymap API for collection
on condition by specifying merge_columns paramterUnified execution engine for both ML and DataFrame
Partition objectdf['new_col'] = df.col + 1dropna parameter for value_countsDelay API which support run DataFrame in parallel
DataFrame database backends including postgresql and mysql
func API for invoking the odps internal functionFix job submission problem for v0.6.x
Fix job submission problem for v0.6.x
Nothing published for this version
DataFrame addfillna and dropna for collection
fillna and dropna for collectionDataFrame add more aggregation including moment, skew, kurtosis, and cat.
reshuffle APIodps.ml is available for machine learning
extract_kv, to_kv, melt APImin_max_scale, std_scale, split, append_idMod support for DataFrame sequence
DataFrame pivot and pivot_table API
pivot and pivot_table API\x00 exists with cython optimized codeHotfix performance of tunnel download
- Fix tunnel, use stream to read - Bugfixes
For DataFrame, astype int64 when encountering a bigint partition field
DataFrame map, apply, agg and map_reduce now support third-party pure python library
Fix job submission problem for v0.5.x
Fix job submission problem for v0.5.x
Fix DataFrame bug when scalar is a Chinese StringScalar for ODPS SQL backend
Fix resource reload, use GET instead of HEAD method
fix switch bug for DataFrame pandas backend
DataFrame SQL backend pypy support
add unknown_as_string and as_type parameters when initializing a DataFrame from pandas.DataFrame
unknown_as_string and as_type parameters when initializing a DataFrame from pandas.DataFrameBetter tunnel performance(cython required)
C integration for Tunnel Download which has better performance
Nothing published for this version
User defined aggregation support
map, apply, agg and map_reduce can read resources from file, table or collectiongraphvizsample APIstrptime for string operationbloom_filter and hll_count APIBuilinFunction for DataFrame to call the builtin function of ODPS SQL which has no parameter or constant parameterssupport aliases for execute_sql
Nothing published for this version
Nothing published for this version
fix bug when creating a table from existed table's schema
Nothing published for this version
Fix bug of MapReduce API, add ascending parameter.
Fix bug when groupby a join expr.
Fix compatible issue under Py3 when running DataFrame API
DataFrame API now support pandas backend
join, union and isin operationsapply on axis=1Nothing published for this version
Nothing published for this version
Your coding agent can read these notes before it upgrades. Set up the MCP server →