pca_sparse_train issue

[Michal Jaslan <ja...@gmail.com>]

Hi,



Has anybody experienced a problem with execution of
madlib.pca_sparse_train function (madlib 1.9.1)?



I faced below exception:

[XX000] ERROR: plpy.SPIError: plpy.SPIError: compresstype "quicklz" is
not supported (plpython.c:4663)

  Gdzie: Traceback (most recent call last):

  PL/Python function "pca_sparse_train", line 23, in <module>

    return pca.pca_sparse(**globals())

  PL/Python function "pca_sparse_train", line 546, in pca_sparse

PL/Python function "pca_sparse_train"



Seems that Madlib expects that HAWQ (2.0.1) I am currently using,
supports quicklz compression alg. Which is not supported by HAWQ.



I would appreciate any hint.



Thanks



Cheers

Michal

Re: pca_sparse_train issue

[Frank McQuillan <fm...@pivotal.io>]

Hi,

Pivotal HDB/HAWQ 2.0.1 does not support QuickLZ compression so you must run
a script to fix that before you can use MADlib.

Details are at:
https://cwiki.apache.org/confluence/display/MADLIB/Installation+Guide#InstallationGuide-10/7/16-InstallationofMADlibonHDB/HAWQ2.0.1


The script is downloadable from:
https://network.pivotal.io/products/pivotal-hdb#/releases/2397/file_groups/487

Please let me know if that does not work for you.

Frank

On Tue, Oct 25, 2016 at 1:11 AM, Michal Jaslan <ja...@gmail.com> wrote:

> Hi,
>
>
>
> Has anybody experienced a problem with execution of madlib.pca_sparse_train function (madlib 1.9.1)?
>
>
>
> I faced below exception:
>
> [XX000] ERROR: plpy.SPIError: plpy.SPIError: compresstype "quicklz" is not supported (plpython.c:4663)
>
>   Gdzie: Traceback (most recent call last):
>
>   PL/Python function "pca_sparse_train", line 23, in <module>
>
>     return pca.pca_sparse(**globals())
>
>   PL/Python function "pca_sparse_train", line 546, in pca_sparse
>
> PL/Python function "pca_sparse_train"
>
>
>
> Seems that Madlib expects that HAWQ (2.0.1) I am currently using, supports quicklz compression alg. Which is not supported by HAWQ.
>
>
>
> I would appreciate any hint.
>
>
>
> Thanks
>
>
>
> Cheers
>
> Michal
>
>