spark-df-profiling
spark-df-profiling copied to clipboard
AttributeError: 'AxesSubplot' object has no attribute 'set_axis_bgcolor'
Running the latest version of this library on EMR using python 3.6.2. Getting the error
AttributeError: 'AxesSubplot' object has no attribute 'set_axis_bgcolor'
. Looks like the library 'matplotlib` had a breaking change as discussed here: https://github.com/scikit-learn/scikit-learn/issues/10762
Same for me. It looks like 'set_axis_bgcolor' should be replaced with 'facecolor'. Please look into part where the histograms are built.
Please change this
File "/usr/local/lib/python2.7/site-packages/spark_df_profiling/base.py", line 164, in mini_histogram plot.set_facecolor("w")
in this line and please add the changes. referer this issue where its changed. - > https://github.com/scikit-learn/scikit-learn/issues/10762
@pidugusundeep I don't follow your comment, this change has been submitted in #14
Thanks @caddac anyways i changed it as i didnt have it when i cloned the repo
@julioasotodv Any chance v1.1.14 will be released soon (including the fix for this issue, hopefully)?
@julioasotodv Waiting for this deploy too, or at least some way to install from source on Azure Databricks
@jeffkpayne just worked around it installing the lib as an egg using this tutorial, but its VERY SLOW, took 36 min to process a (600k, 50) dataframe on a small cluster, converting it to pandas then using pandas profiling is taking about 1-2 min
Any PRs to address this issue? Not working in Azure Databricks with python 3.6