| Commit message (Collapse) | Author | Age | Files | Lines |
... | |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
provided schema
When creating a DataFrame from an RDD in PySpark, `createDataFrame` calls `.take(10)` to verify the first 10 rows of the RDD match the provided schema. Similar to https://issues.apache.org/jira/browse/SPARK-8070, but that issue affected cases where a schema was not provided.
Verifying the first 10 rows is of limited utility and causes the DAG to be executed non-lazily. If necessary, I believe this verification should be done lazily on all rows. However, since the caller is providing a schema to follow, I think it's acceptable to simply fail if the schema is incorrect.
marmbrus We chatted about this at SparkSummitEU. davies you made a similar change for the infer-schema path in https://github.com/apache/spark/pull/6606
Author: Jason White <jason.white@shopify.com>
Closes #9392 from JasonMWhite/createDataFrame_without_take.
|
|
|
|
|
|
|
|
|
|
| |
JIRA: https://issues.apache.org/jira/browse/SPARK-11322
As reported by JoshRosen in [databricks/spark-redshift/issues/89](https://github.com/databricks/spark-redshift/issues/89#issuecomment-149828308), the exception-masking behavior sometimes makes debugging harder. To deal with this issue, we should keep full stack trace in the captured exception.
Author: Liang-Chi Hsieh <viirya@appier.com>
Closes #9283 from viirya/py-exception-stacktrace.
|
|
|
|
|
|
|
|
| |
Adds DataFrameReader.text and DataFrameWriter.text.
Author: Reynold Xin <rxin@databricks.com>
Closes #9259 from rxin/SPARK-11292.
|
|
|
|
|
|
|
|
| |
[SPARK-10668](https://issues.apache.org/jira/browse/SPARK-10668) has provided ```WeightedLeastSquares``` solver("normal") in ```LinearRegression``` with L2 regularization in Scala and R, Python ML ```LinearRegression``` should also support setting solver("auto", "normal", "l-bfgs")
Author: Yanbo Liang <ybliang8@gmail.com>
Closes #9328 from yanboliang/spark-11367.
|
|
|
|
|
|
|
|
|
|
|
|
| |
returns incorrect answer in some cases
Fix computation of root-sigma-inverse in multivariate Gaussian; add a test and fix related Python mixture model test.
Supersedes https://github.com/apache/spark/pull/9293
Author: Sean Owen <sowen@cloudera.com>
Closes #9309 from srowen/SPARK-11302.2.
|
|
|
|
|
|
|
|
|
| |
implement {RandomForest, GBT, TreeEnsemble, TreeClassifier, TreeRegressor}Params for Python API
in pyspark/ml/{classification, regression}.py
Author: vectorijk <jiangkai@gmail.com>
Closes #9233 from vectorijk/spark-10024.
|
|
|
|
|
|
|
|
|
|
| |
BlockMatrix
This PR adds addition and multiplication to PySpark's `BlockMatrix` class via `add` and `multiply` functions.
Author: Mike Dusenberry <mwdusenb@us.ibm.com>
Closes #9139 from dusenberrymw/SPARK-6488_Add_Addition_and_Multiplication_to_PySpark_BlockMatrix.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
from the Kafka Streaming API
jerryshao tdas
I know this is kind of minor, and I know you all are busy, but this brings this class in line with the `OffsetRange` class, and makes tests a little more concise.
Instead of doing something like:
```
assert topic_and_partition_instance._topic == "foo"
assert topic_and_partition_instance._partition == 0
```
You can do something like:
```
assert topic_and_partition_instance == TopicAndPartition("foo", 0)
```
Before:
```
>>> from pyspark.streaming.kafka import TopicAndPartition
>>> TopicAndPartition("foo", 0) == TopicAndPartition("foo", 0)
False
```
After:
```
>>> from pyspark.streaming.kafka import TopicAndPartition
>>> TopicAndPartition("foo", 0) == TopicAndPartition("foo", 0)
True
```
I couldn't find any tests - am I missing something?
Author: Nick Evans <me@nicolasevans.org>
Closes #9236 from manygrams/topic_and_partition_equality.
|
|
|
|
|
|
|
|
|
|
| |
Duplicated the since decorator from pyspark.sql into pyspark (also tweaked to handle functions without docstrings).
Added since to methods + "versionadded::" to classes (derived from the git file history in pyspark).
Author: noelsmith <mail@noelsmith.com>
Closes #8627 from noel-smith/SPARK-10271-since-mllib-clustering.
|
|
|
|
|
|
| |
Author: Jeff Zhang <zjffdu@apache.org>
Closes #9248 from zjffdu/SPARK-11279.
|
|
|
|
|
|
|
|
| |
pyspark.mllib.regression
Author: Yu ISHIKAWA <yuu.ishikawa@gmail.com>
Closes #8684 from yu-iskw/SPARK-10277.
|
|
|
|
|
|
|
|
| |
WIP
Author: Gábor Lipták <gliptak@gmail.com>
Closes #8323 from gliptak/SPARK-7021.
|
|
|
|
|
|
|
|
|
|
| |
…rint in python
No test needed. Verify it manually in pyspark shell
Author: Jeff Zhang <zjffdu@apache.org>
Closes #9177 from zjffdu/SPARK-11205.
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Without an empty line, sphinx will treat doctest as docstring. holdenk
~~~
/Users/meng/src/spark/python/pyspark/ml/feature.py:docstring of pyspark.ml.feature.CountVectorizer:3: ERROR: Undefined substitution referenced: "label|raw |vectors | +-----+---------------+-------------------------+ |0 |[a, b, c] |(3,[0,1,2],[1.0,1.0,1.0])".
/Users/meng/src/spark/python/pyspark/ml/feature.py:docstring of pyspark.ml.feature.CountVectorizer:3: ERROR: Undefined substitution referenced: "1 |[a, b, b, c, a]|(3,[0,1,2],[2.0,2.0,1.0])".
~~~
Author: Xiangrui Meng <meng@databricks.com>
Closes #9188 from mengxr/py-count-vec-doc-fix.
|
|
|
|
|
|
|
|
| |
Namely "." shows up in some places in the template when using the param docstring and not in others
Author: Holden Karau <holden@pigscanfly.ca>
Closes #9017 from holdenk/SPARK-10767-Make-pyspark-shared-params-codegen-more-consistent.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
pyspark.mllib.classification
Duplicated the since decorator from pyspark.sql into pyspark (also tweaked to handle functions without docstrings).
Added since to methods + "versionadded::" to classes derived from the file history.
Note - some methods are inherited from the regression module (i.e. LinearModel.intercept) so these won't have version numbers in the API docs until that model is updated.
Author: noelsmith <mail@noelsmith.com>
Closes #8626 from noel-smith/SPARK-10269-since-mlib-classification.
|
|
|
|
|
|
|
|
|
|
|
|
| |
Duplicated the since decorator from pyspark.sql into pyspark (also tweaked to handle functions without docstrings).
Added since to public methods + "versionadded::" to classes (derived from the git file history in pyspark).
Note - I added also the tags to MultilabelMetrics even though it isn't declared as public in the __all__ statement... if that's incorrect - I'll remove.
Author: noelsmith <mail@noelsmith.com>
Closes #8628 from noel-smith/SPARK-10272-since-mllib-evalutation.
|
|
|
|
|
|
|
|
|
| |
Upgrade to Py4j0.9
Author: Holden Karau <holden@pigscanfly.ca>
Author: Holden Karau <holden@us.ibm.com>
Closes #8615 from holdenk/SPARK-10447-upgrade-pyspark-to-py4j0.9.
|
|
|
|
|
|
|
|
| |
Also added SQLContext.newSession()
Author: Davies Liu <davies@databricks.com>
Closes #9122 from davies/py_create.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This commit refactors the `run-tests-jenkins` script into Python. This refactoring was done by brennonyork in #7401; this PR contains a few minor edits from joshrosen in order to bring it up to date with other recent changes.
From the original PR description (by brennonyork):
Currently a few things are left out that, could and I think should, be smaller JIRA's after this.
1. There are still a few areas where we use environment variables where we don't need to (like `CURRENT_BLOCK`). I might get around to fixing this one in lieu of everything else, but wanted to point that out.
2. The PR tests are still written in bash. I opted to not change those and just rewrite the runner into Python. This is a great follow-on JIRA IMO.
3. All of the linting scripts are still in bash as well and would likely do to just add those in as follow-on JIRA's as well.
Closes #7401.
Author: Brennon York <brennon.york@capitalone.com>
Closes #9161 from JoshRosen/run-tests-jenkins-refactoring.
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
by presenting the Object
The _verify_type() function had Errors that were raised when there were Type conversion issues but left out the Object in question. The Object is now added in the Error to reduce the strain on the user to debug through to figure out the Object that failed the Type conversion.
The use case for me was a Pandas DataFrame that contained 'nan' as values for columns of Strings.
Author: Mahmoud Lababidi <mahmoud@thehumangeo.com>
Author: Mahmoud Lababidi <lababidi@gmail.com>
Closes #9149 from lababidi/master.
|
|
|
|
|
|
|
|
| |
Make sure comma-separated paths get processed correcly in ResolvedDataSource for a HadoopFsRelationProvider
Author: Koert Kuipers <koert@tresata.com>
Closes #8416 from koertkuipers/feat-sql-comma-separated-paths.
|
|
|
|
|
|
|
|
|
|
| |
binary search
At this moment `SparseVector.__getitem__` executes `np.searchsorted` first and checks if result is in an expected range after that. It is possible to check if index can contain non-zero value before executing `np.searchsorted`.
Author: zero323 <matthew.szymkiewicz@gmail.com>
Closes #9098 from zero323/sparse_vector_getitem_improved.
|
|
|
|
|
|
|
|
|
|
|
|
| |
…rror message
For negative indices in the SparseVector, we update the index value. If we have an incorrect index
at this point, the error message has the incorrect *updated* index instead of the original one. This
change contains the fix for the same.
Author: Bhargav Mangipudi <bhargav.mangipudi@gmail.com>
Closes #9069 from bhargav/spark-10759.
|
|
|
|
|
|
|
|
|
|
| |
Output list of supported modules for python tests in error message when given bad module name.
CC: davies
Author: Joseph K. Bradley <joseph@databricks.com>
Closes #9088 from jkbradley/python-tests-modules.
|
|
|
|
|
|
|
|
|
|
| |
cancel all running jobs
This patch adds a signal handler to trap Ctrl-C and cancels running job.
Author: Ashwin Shankar <ashankar@netflix.com>
Closes #9033 from ashwinshankar77/master.
|
|
|
|
|
|
|
|
|
|
| |
PySpark
Support for recommendUsersForProducts and recommendProductsForUsers in matrix factorization model for PySpark
Author: Vladimir Vladimirov <vladimir.vladimirov@magnetic.com>
Closes #8700 from smartkiwi/SPARK-10535_.
|
|
|
|
|
|
|
|
|
|
| |
with given regParam and convergenceTol parameters
These params were being passed into the StreamingLogisticRegressionWithSGD constructor, but not transferred to the call for model training. Same with StreamingLinearRegressionWithSGD. I added the params as named arguments to the call and also fixed the intercept parameter, which was being passed as regularization value.
Author: Bryan Cutler <bjcutler@us.ibm.com>
Closes #9002 from BryanCutler/StreamingSGD-convergenceTol-bug-10959.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
when we…
__gettitem__ method throws IndexError exception when we try to access index after the last non-zero entry
from pyspark.mllib.linalg import Vectors
sv = Vectors.sparse(5, {1: 3})
sv[0]
## 0.0
sv[1]
## 3.0
sv[2]
## Traceback (most recent call last):
## File "<stdin>", line 1, in <module>
## File "/python/pyspark/mllib/linalg/__init__.py", line 734, in __getitem__
## row_ind = inds[insert_index]
## IndexError: index out of bounds
Author: zero323 <matthew.szymkiewicz@gmail.com>
Closes #9009 from zero323/sparse_vector_index_error.
|
|
|
|
|
|
|
|
| |
Add the Python API for isotonicregression.
Author: Holden Karau <holden@pigscanfly.ca>
Closes #8214 from holdenk/SPARK-9774-add-python-api-for-ml-regression-isotonicregression.
|
|
|
|
|
|
|
|
|
|
| |
(spark.mllib)
Provide initialModel param for pyspark.mllib.clustering.KMeans
Author: Evan Chen <chene@us.ibm.com>
Closes #8967 from evanyc15/SPARK-10779-pyspark-mllib.
|
|
|
|
|
|
|
|
|
|
| |
PySpark's AFTSurvivalRegression
If user doesn't specify `quantileProbs` in `setParams`, it will get reset to the default value. We don't need special handling here. vectorijk yanboliang
Author: Xiangrui Meng <meng@databricks.com>
Closes #9001 from mengxr/SPARK-10957.
|
|
|
|
|
|
|
|
| |
Implement Python API for AFTSurvivalRegression
Author: vectorijk <jiangkai@gmail.com>
Closes #8926 from vectorijk/spark-10688.
|
|
|
|
|
|
|
|
| |
Documentation for dropDuplicates() and drop_duplicates() is one and the same. Resolved the error in the example for drop_duplicates using the same approach used for groupby and groupBy, by indicating that dropDuplicates and drop_duplicates are aliases.
Author: asokadiggs <asoka.diggs@intel.com>
Closes #8930 from asokadiggs/jira-10782.
|
|
|
|
|
|
|
|
|
|
|
|
| |
Add method to easily convert a StatCounter instance into a Python dict
https://issues.apache.org/jira/browse/SPARK-6919
Note: This is my original work and the existing Spark license applies.
Author: Erik Shilts <erik.shilts@opower.com>
Closes #5516 from eshilts/statcounter-asdict.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
These are CSS/JavaScript changes changes to make navigation in the PySpark API a bit simpler by adding the following to the sidebar:
* Classes
* Functions
* Tags to highlight experimental features
![screen shot 2015-09-02 at 08 50 12](https://cloud.githubusercontent.com/assets/11915197/9634781/301f853a-518b-11e5-8d5c-fda202f6202f.png)
Online example here: https://dl.dropboxusercontent.com/u/20821334/pyspark-api-nav-enhance/pyspark.mllib.html
(The contribution is my original work and that I license the work to the project under the project's open source license)
Author: noelsmith <mail@noelsmith.com>
Closes #8571 from noel-smith/pyspark-api-nav-enhance.
|
|
|
|
|
|
|
|
|
|
|
|
| |
This integrates the Interaction feature transformer with SparkR R formula support (i.e. support `:`).
To generate reasonable ML attribute names for feature interactions, it was necessary to add the ability to read attribute the original attribute names back from `StructField`, and also to specify custom group prefixes in `VectorAssembler`. This also has the side-benefit of cleaning up the double-underscores in the attributes generated for non-interaction terms.
mengxr
Author: Eric Liang <ekl@databricks.com>
Closes #8830 from ericl/interaction-2.
|
|
|
|
|
|
|
|
|
|
|
|
| |
Python DataFrame.
Python DataFrame.head/take now requires scanning all the partitions. This pull request changes them to delegate the actual implementation to Scala DataFrame (by calling DataFrame.take).
This is more of a hack for fixing this issue in 1.5.1. A more proper fix is to change executeCollect and executeTake to return InternalRow rather than Row, and thus eliminate the extra round-trip conversion.
Author: Reynold Xin <rxin@databricks.com>
Closes #8876 from rxin/SPARK-10731.
|
|
|
|
|
|
|
|
|
|
|
|
| |
usingColumns
JIRA: https://issues.apache.org/jira/browse/SPARK-10446
Currently the method `join(right: DataFrame, usingColumns: Seq[String])` only supports inner join. It is more convenient to have it support other join types.
Author: Liang-Chi Hsieh <viirya@appier.com>
Closes #8600 from viirya/usingcolumns_df.
|
|
|
|
|
|
|
|
| |
https://issues.apache.org/jira/browse/SPARK-10577
Author: Jian Feng <jzhang.chs@gmail.com>
Closes #8801 from Jianfeng-chs/master.
|
|
|
|
|
|
|
|
|
|
| |
on OS X due to hidden file
Remove ._SUCCESS.crc hidden file that may cause problems in distribution tar archive, and is not used
Author: Sean Owen <sowen@cloudera.com>
Closes #8846 from srowen/SPARK-10716.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
from the issue:
In Scala, I can supply a custom partitioner to reduceByKey (and other aggregation/repartitioning methods like aggregateByKey and combinedByKey), but as far as I can tell from the Pyspark API, there's no way to do the same in Python.
Here's an example of my code in Scala:
weblogs.map(s => (getFileType(s), 1)).reduceByKey(new FileTypePartitioner(),_+_)
But I can't figure out how to do the same in Python. The closest I can get is to call repartition before reduceByKey like so:
weblogs.map(lambda s: (getFileType(s), 1)).partitionBy(3,hash_filetype).reduceByKey(lambda v1,v2: v1+v2).collect()
But that defeats the purpose, because I'm shuffling twice instead of once, so my performance is worse instead of better.
Author: Holden Karau <holden@pigscanfly.ca>
Closes #8569 from holdenk/SPARK-9821-pyspark-reduceByKey-should-take-a-custom-partitioner.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Added newlines before `:param ...:` and `:return:` markup. Without these, parameter lists aren't formatted correctly in the API docs. I.e:
![screen shot 2015-09-21 at 21 49 26](https://cloud.githubusercontent.com/assets/11915197/10004686/de3c41d4-60aa-11e5-9c50-a46dcb51243f.png)
.. looks like this once newline is added:
![screen shot 2015-09-21 at 21 50 14](https://cloud.githubusercontent.com/assets/11915197/10004706/f86bfb08-60aa-11e5-8524-ae4436713502.png)
Author: noelsmith <mail@noelsmith.com>
Closes #8851 from noel-smith/docstring-missing-newline-fix.
|
|
|
|
|
|
|
|
| |
From JIRA: Add Python API, user guide and example for ml.feature.CountVectorizerModel
Author: Holden Karau <holden@pigscanfly.ca>
Closes #8561 from holdenk/SPARK-9769-add-python-api-for-countvectorizermodel.
|
|
|
|
|
|
|
|
| |
There are some missing API docs in pyspark.mllib.linalg.Vector (including DenseVector and SparseVector). We should add them based on their Scala counterparts.
Author: vinodkc <vinod.kc.in@gmail.com>
Closes #8834 from vinodkc/fix_SPARK-10631.
|
|
|
|
|
|
|
|
|
|
| |
It does not make much sense to set `spark.shuffle.spill` or `spark.sql.planner.externalSort` to false: I believe that these configurations were initially added as "escape hatches" to guard against bugs in the external operators, but these operators are now mature and well-tested. In addition, these configurations are not handled in a consistent way anymore: SQL's Tungsten codepath ignores these configurations and will continue to use spilling operators. Similarly, Spark Core's `tungsten-sort` shuffle manager does not respect `spark.shuffle.spill=false`.
This pull request removes these configurations, adds warnings at the appropriate places, and deletes a large amount of code which was only used in code paths that did not support spilling.
Author: Josh Rosen <joshrosen@databricks.com>
Closes #8831 from JoshRosen/remove-ability-to-disable-spilling.
|
|
|
|
|
|
|
|
| |
As ```assertEquals``` is deprecated, so we need to change ```assertEquals``` to ```assertEqual``` for existing python unit tests.
Author: Yanbo Liang <ybliang8@gmail.com>
Closes #8814 from yanboliang/spark-10615.
|
|
|
|
|
|
|
|
|
|
| |
JIRA: https://issues.apache.org/jira/browse/SPARK-10642
When calling `rdd.lookup()` on a RDD with tuple keys, `portable_hash` will return a long. That causes `DAGScheduler.submitJob` to throw `java.lang.ClassCastException: java.lang.Long cannot be cast to java.lang.Integer`.
Author: Liang-Chi Hsieh <viirya@appier.com>
Closes #8796 from viirya/fix-pyrdd-lookup.
|
|
|
|
|
|
|
|
| |
pyspark.ml.recommendation
Author: Yu ISHIKAWA <yuu.ishikawa@gmail.com>
Closes #8692 from yu-iskw/SPARK-10282.
|
|
|
|
|
|
| |
Author: Yu ISHIKAWA <yuu.ishikawa@gmail.com>
Closes #8665 from yu-iskw/SPARK-10274.
|