spark-df-profiling icon indicating copy to clipboard operation
spark-df-profiling copied to clipboard

'DataFrame' object has no attribute 'ix'

Open rubenssoto opened this issue 4 years ago • 8 comments

I installed by pip, when i try yo profilling my dataframe this errors appers

'DataFrame' object has no attribute 'ix'

Thank you

rubenssoto avatar Apr 07 '20 23:04 rubenssoto

Same here. I'm using pandas=1.0.3.

felipeeeantunes avatar Apr 28 '20 19:04 felipeeeantunes

Same here, it is because 'ix' has been removed for pandas > 1.x.x, in code it is results_data.ix[0] and it must be change to iloc. I'm using pandas v1.0.3 as well.

crar01 avatar May 24 '20 17:05 crar01

I just ran into this problem as well. I specified pandas version 0.25.1 which seems to have worked.

For reference here are the dependency versions I ran with successfully in PyCharm

  • Python 3.7
  • Apache Spark 2.4.6
  • An internet connection
  • jinja2 2.11.2
  • matplotlib 3.2.2
  • pandas 0.25.1
  • six 1.15.0

mrperson2015 avatar Jul 03 '20 00:07 mrperson2015

Correction to the above:

  • Python 3.7
  • Apache Spark 2.4.6
  • An internet connection
  • jinja2 2.11.2
  • matplotlib 1.5.3
  • pandas 0.25.1
  • six 1.15.0

leocasarsa avatar Jul 20 '20 15:07 leocasarsa

Just in case anyone else comes across this, the current version on GitHub solves the .ix problem after the work in #36 but the release hasn't been published as mentioned in #33 ;

I comment on that issue how to use pip to pull directly the Git version if anyone goes down that route.

BrunoGomesCoelho avatar Aug 18 '20 15:08 BrunoGomesCoelho

Any other fix on this outside of downgrading pandas?

My pandas won't revert to 0.25.1

StevenPierre avatar Nov 03 '22 19:11 StevenPierre

Its quite hard to change packages down to such an old version these days and the error still exists for me..

jbogensperger avatar Dec 04 '23 10:12 jbogensperger

You can install directly from the latest version that has it fixed, it was just never released - For how to add that to a requirements.txt see my comment in #33 or use the same comment for pip (e.g. pip install git+https://github.com/julioasotodv/spark-df-profiling)

BrunoGomesCoelho avatar Dec 04 '23 17:12 BrunoGomesCoelho