pyspark 2.4 support

[mattomatic]

Does this library currently work with spark 2.4?

[icexelloss]

We have not tried it with Spark 2.4 yet.

…

On Wed, Feb 13, 2019 at 3:44 PM mattomatic ***@***.***> wrote: Does this library currently work with spark 2.4? — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#63>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAwbrHasxpH0N4PydNjQOOvfgBQU4YFJks5vNHkagaJpZM4a6PJ7> .

[andompesta]

It does not work with 2.3.2 as well

[icexelloss]

what issues do you see with 2.3.2? Internally we use flint with 2.3.2 without issues.

…

On Fri, Feb 15, 2019 at 1:53 AM Sandro Cavallari ***@***.***> wrote: It does not work wit 2.3.2 as well — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#63 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAwbrIZKDz7cGE8o1d7di37jFP3jI6yCks5vNlmFgaJpZM4a6PJ7> .

[mattomatic]

#64

Managed these changes to get it to build under spark 2.4

[kunalsingh09]

I am running on spark-2.4.0-bin-hadoop2.7 and seeing the below error. Any idea ?

from ts.flint import windows
sp500_previous_day_return = sp500_return.shiftTime(windows.future_absolute_time('1day')).toDF('time', 'previous_day_return')
Traceback (most recent call last):
File "", line 1, in
File "/Users/kkum25/anaconda/envs/featuretool/lib/python3.7/site-packages/ts/flint/dataframe.py", line 1591, in shiftTime
tsrdd = self.timeSeriesRDD.shift(window._jwindow(self._sc))
File "/Users/kkum25/anaconda/envs/featuretool/lib/python3.7/site-packages/ts/flint/dataframe.py", line 154, in timeSeriesRDD
self._jdf, self._is_sorted, self._junit, self._time_column)
File "/usr/local/Cellar/apache-spark/2.2.1/libexec/python/lib/py4j-0.10.4-src.zip/py4j/java_gateway.py", line 1133, in call
File "/usr/local/Cellar/apache-spark/2.2.1/libexec/python/pyspark/sql/utils.py", line 63, in deco
return f(*a, **kw)
File "/usr/local/Cellar/apache-spark/2.2.1/libexec/python/lib/py4j-0.10.4-src.zip/py4j/protocol.py", line 319, in get_return_value
py4j.protocol.Py4JJavaError: An error occurred while calling o124.fromDF.
: java.lang.NoSuchMethodError: org.apache.spark.sql.catalyst.plans.physical.ClusteredDistribution$.apply$default$2()Lscala/Option;
at com.twosigma.flint.timeseries.TimeSeriesStore$.isClustered(TimeSeriesStore.scala:149)
at com.twosigma.flint.timeseries.TimeSeriesStore$.apply(TimeSeriesStore.scala:64)
at com.twosigma.flint.timeseries.TimeSeriesRDD$.fromDFWithPartInfo(TimeSeriesRDD.scala:509)
at com.twosigma.flint.timeseries.TimeSeriesRDD$.fromDF(TimeSeriesRDD.scala:304)
at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
at sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:62)
at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
at java.lang.reflect.Method.invoke(Method.java:498)
at py4j.reflection.MethodInvoker.invoke(MethodInvoker.java:244)
at py4j.reflection.ReflectionEngine.invoke(ReflectionEngine.java:357)
at py4j.Gateway.invoke(Gateway.java:280)
at py4j.commands.AbstractCommand.invokeMethod(AbstractCommand.java:132)
at py4j.commands.CallCommand.execute(CallCommand.java:79)
at py4j.GatewayConnection.run(GatewayConnection.java:214)
at java.lang.Thread.run(Thread.java:745)

[kunalsingh09]

#64

Managed these changes to get it to build under spark 2.4

before this change what issue you were seeing ?

[luzlab]

Is there any blockers to merging this in? I'd like to use Flint on Databricks but I don't see any compatible versions of Spark being offered (2.2.1 or 2.4.[0,1,2] are only versions of Spark currently available).

I think this is a great project and would love to help mature it!

[dgrnbrg]

I have been successfully using @mattomatic's changes to run on Spark 2.4.

[prabhash17]

@icexelloss Does this library work with spark 2.2.x version?