pytables-users

[Pytables-users] Questionnaire

[Francesc A. <fa...@ca...>]

Hi List,

Now that PyTables has passed its second birthday, we need your feedback
again. So, please, if you are using PyTables and like this effort to be
continued and improved, take some time to respond to this
questionnaire; it will help us to decide what to do in the next few
months.

=46ollowing are our plans for the next few months, before releasing the
1.0 version. Could you respond to these questions by giving a
punctuation ranging from +5 (I absolutely need that) to 0 (I can pass
without this) to these planned features?

=2D Implement support for Variable Length values in Table columns (most
  likely implemented in a brand-new object called VLTables).

=2D Implement relationships (apart from the existing hierarchical ones)
  between objects in the object tree (kind of symbolic links on a
  filesystems), and pointers to dataset regions.

=2D Implement date support in Leaf objects (specially on Tables).

=2D Full index-based searches in Tables (right now, you can only set one
  condition in index-accelerated searches).

=2D Image support (in the sense stated in
  http://hdf.ncsa.uiuc.edu/HDF5/doc/ADGuide/ImageSpec.html).


Now, answer the next questions. If you don't like to respond all of them,
this is not necessary.

=2D What feature do you like more (not necessarily listed before)?

=2D What feature do you miss more (not necessarily listed before)?

=2D Do you prefer seeing PyTables to become (even) faster or less memory
  demanding?

=2D Which present limitations do you find more annoying?

=2D In which field of engineering, science or business are you using
  PyTables?


Ok, that's all. I hope some of you will take some of his precious time
to fill-up the questionnaire. Even only one answer would be much
better than our sole opinion!.

Thanks,

=2D-=20
=46rancesc Altet
Who's your data daddy? =A0PyTables

Re: [Pytables-users] Questionnaire

[Andrew S. <str...@as...>]

Given that I'm an "interested and intermittent" user, not a "hardcore, 
must have" user, please don't rank my responses as highly as others'.  
Also, I've only answered the questions that I feel even remotely 
qualified or interested to answer.  (You may take my other answers as "0".)

>- Image support (in the sense stated in
>  http://hdf.ncsa.uiuc.edu/HDF5/doc/ADGuide/ImageSpec.html).
>  
>
Sounds great
+1

In conjunction with releasing Python GIL on file write operations: +5

This might make me a hardcore user if I convert my primitive home-brew 
"movie" code. But this would require releasing the Python GIL on writing 
the data to HDF5 files.  This would allow a multi-threaded Python 
program to continue working while the (non-Python API using) C code 
completes. Refer to our previous conversation: 
http://sourceforge.net/mailarchive/forum.php?thread_id=4963045&forum_id=13760 
My email (at sourceforge time 2004-06-22 00:04) describes the (seemingly 
fairly simple) necessary changes.  (Sorry, if you've done this already 
-- I haven't kept up with the changes in the last few months.)

>- Full index-based searches in Tables (right now, you can only set one
>  condition in index-accelerated searches).
>  
>
Sounds good, but not informed enough to judge (+1?)

>- Which present limitations do you find more annoying?
>  
>
(Sorry to keep kicking a dead horse.) Keeping the GIL in long file write 
operations where the Python API is not touched and therefore it could be 
released.

>- In which field of engineering, science or business are you using
>  PyTables?
>  
>
Neuroscience.

Re: [Pytables-users] Questionnaire

[Francesc A. <fa...@ca...>]

Hi Andrew,

Yes, releasing the GIL when writing (or reading) data would be a very nice
addition. I've to look more in depth in that. That would be hopefully
included in version 1.0 of PyTables.

Anyway, if you know how to do add that kind of support, and test it, I'll be
glad to include your code in next version of PyTables.

Thanks for your time,

A Dimarts 30 Novembre 2004 21:18, Andrew Straw va escriure:
> Given that I'm an "interested and intermittent" user, not a "hardcore,=20
> must have" user, please don't rank my responses as highly as others'. =20
> Also, I've only answered the questions that I feel even remotely=20
> qualified or interested to answer.  (You may take my other answers as "0"=
=2E)
>=20
> >- Image support (in the sense stated in
> >  http://hdf.ncsa.uiuc.edu/HDF5/doc/ADGuide/ImageSpec.html).
> > =20
> >
> Sounds great
> +1
>=20
> In conjunction with releasing Python GIL on file write operations: +5
>=20
> This might make me a hardcore user if I convert my primitive home-brew=20
> "movie" code. But this would require releasing the Python GIL on writing=
=20
> the data to HDF5 files.  This would allow a multi-threaded Python=20
> program to continue working while the (non-Python API using) C code=20
> completes. Refer to our previous conversation:=20
> http://sourceforge.net/mailarchive/forum.php?thread_id=3D4963045&forum_id=
=3D13760=20
> My email (at sourceforge time 2004-06-22 00:04) describes the (seemingly=
=20
> fairly simple) necessary changes.  (Sorry, if you've done this already=20
> -- I haven't kept up with the changes in the last few months.)
>=20
> >- Full index-based searches in Tables (right now, you can only set one
> >  condition in index-accelerated searches).
> > =20
> >
> Sounds good, but not informed enough to judge (+1?)
>=20
> >- Which present limitations do you find more annoying?
> > =20
> >
> (Sorry to keep kicking a dead horse.) Keeping the GIL in long file write=
=20
> operations where the Python API is not touched and therefore it could be=
=20
> released.
>=20
> >- In which field of engineering, science or business are you using
> >  PyTables?
> > =20
> >
> Neuroscience.
>=20
>=20
>=20
>=20
> -------------------------------------------------------
> SF email is sponsored by - The IT Product Guide
> Read honest & candid reviews on hundreds of IT Products from real users.
> Discover which products truly live up to the hype. Start reading now.=20
> http://productguide.itmanagersjournal.com/
> _______________________________________________
> Pytables-users mailing list
> Pyt...@li...
> https://lists.sourceforge.net/lists/listinfo/pytables-users
>=20

=2D-=20
=46rancesc Altet
Who's your data daddy? =A0PyTables

Re: [Pytables-users] Questionnaire

[Ashley W. <ash...@sy...>]

On 30/11/2004, at 4:16 AM, Francesc Altet wrote:

> Hi List,
>
> Now that PyTables has passed its second birthday, we need your =
feedback
> again. So, please, if you are using PyTables and like this effort to =
be
> continued and improved, take some time to respond to this
> questionnaire; it will help us to decide what to do in the next few
> months.
>
> Following are our plans for the next few months, before releasing the
> 1.0 version. Could you respond to these questions by giving a
> punctuation ranging from +5 (I absolutely need that) to 0 (I can pass
> without this) to these planned features?
>
> - Implement support for Variable Length values in Table columns (most
>   likely implemented in a brand-new object called VLTables).

+2, interesting

> - Implement relationships (apart from the existing hierarchical ones)
>   between objects in the object tree (kind of symbolic links on a
>   filesystems), and pointers to dataset regions.

+2, interesting

> - Implement date support in Leaf objects (specially on Tables).

+5, tables with a time column are our stock in trade.

> - Full index-based searches in Tables (right now, you can only set one
>   condition in index-accelerated searches).

+3, useful, but I haven't done enough searching to know how useful yet.

> - Image support (in the sense stated in
>   http://hdf.ncsa.uiuc.edu/HDF5/doc/ADGuide/ImageSpec.html).

0, haven't stored an image yet

> Now, answer the next questions. If you don't like to respond all of=20
> them,
> this is not necessary.
>
> - What feature do you like more (not necessarily listed before)?

CSTables ;)  Or more to the point to ability to have one write-access=20
point, but easy read-access from other processes.

> - What feature do you miss more (not necessarily listed before)?
>
> - Do you prefer seeing PyTables to become (even) faster or less memory
>   demanding?

Both?

> - Which present limitations do you find more annoying?
>
> - In which field of engineering, science or business are you using
>   PyTables?

Engineering operations management.  Process analysis and optimisation. =20=

Reports for operators and engineers.

> Ok, that's all. I hope some of you will take some of his precious time
> to fill-up the questionnaire. Even only one answer would be much
> better than our sole opinion!.
>
> Thanks,
>
> --=20
> Francesc Altet
> Who's your data daddy? =A0PyTables


Thank you.

Ashley Walsh



PS sourceforge is down at the moment or I'd put this in as a bug.

Table.iterrows() should probably return an empty iterator (eg  "return=20=

iter([])" ) for the case when start >=3D stop.  Returning an empty=20
RecArray raises a TypeError for:

   iter(table)

because iter() expects an iterator to be returned by table.__iter__

Re: [Pytables-users] Questionnaire

[Francesc A. <fa...@ca...>]

Hi Ashley,

Thanks for filling-up the questionnaire. Find some notes intertwined in the
text behind.

A Dijous 09 Desembre 2004 09:20, Ashley Walsh va escriure:
> > - Implement date support in Leaf objects (specially on Tables).
>=20
> +5, tables with a time column are our stock in trade.

Well, I think we will be definitely able to deliver that for 1.0

> > - What feature do you like more (not necessarily listed before)?
>=20
> CSTables ;)  Or more to the point to ability to have one write-access=20
> point, but easy read-access from other processes.

Great, perhaps you will be interested in testing the beta version that we
will be releasing by the end of the month. Stay tuned.

> PS sourceforge is down at the moment or I'd put this in as a bug.
>=20
> Table.iterrows() should probably return an empty iterator (eg  "return=20
> iter([])" ) for the case when start >=3D stop.  Returning an empty=20
> RecArray raises a TypeError for:
>=20
>    iter(table)
>=20
> because iter() expects an iterator to be returned by table.__iter__

Good point. This is implemented in CVS (MAIN trunk) now.

Cheers,

=2D-=20
=46rancesc Altet
Who's your data daddy? =A0PyTables