Issues Fixed in Cloudera Data Science Workbench 1.6.0
The current release of Cloudera Data Science Workbench includes fixes for bugs.
-
Fixed an issue where you had to include
pd.options.display.html.table_schema = True
to show a horizontal scroll bar for Pandas Dataframe if there were too many columns. You no longer have to include the property.Cloudera Bug: DSE-3562
- Fixed an issue where the built-in Workbench editor did not properly recognize
imported code that uses tabs instead of spaces. This also resolves navigation issues
that occurred within the editor when working with imported code that uses
tabs.
Cloudera Bug: DSE-2976, DSE-3221
-
Fixed an issue where an email with attachments triggered by a job fail to send if the attachment is over 4 MB.
Cloudera Bug: DSE-5980, DSE-6003
-
Fixed an issue where large R scripts do not respond when run in the built-in Workbench editor.
Cloudera Bug: DSE-2817
-
Fixed an issue where
.md
files were not rendered in Markdown. Previously, onlyREADME.md
was rendered correctly.Cloudera Bug: DSE-3315
-
Fixed an issue with
predict.py
, the model training script in the Python template project.Cloudera Bug: DSE-5314
-
Fixed an issue where logs generated by the Cloudera Data Science Workbench diagnostic bundle were occupying too much space the
/var/log/cdsw
directory. The size of the generated bundle has been reduced and you can now configure a temporary staging directory to be used when a diagnostic bundle is generated.Cloudera Bug: DSE-5921
-
The
cdsw-build.sh
script used with models and experiments now runs as thecdsw
user.Cloudera Bug: DSE-4340
-
The changes to GPU support in version 1.6 have also fixed an issue where GPUs were not automatically detected after a machine reboot.
Cloudera Bug: DSE-2847
-
Fixed an issue where iFrame visualizations would not render in the Workbench due to the new HTTP security headers added in version 1.4.x.
Cloudera Bug: DSE-5274