Miki Tebeka | 3 Feb 06:10 2015
Picon

[Q] Units library?

Greetings,

Any recommendations for a library/data type the keep track of units (kg, inch...). I found several when googling and would like to hear some opinions.

Thanks,
--
Miki

--
You received this message because you are subscribed to the Google Groups "PyData" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pydata+unsubscribe-/JYPxA39Uh5TLH3MbocFFw@public.gmane.org.
For more options, visit https://groups.google.com/d/optout.
Adam | 3 Feb 02:13 2015
Picon

Monkey Patching Panel.__getitem__()

I'm trying to monkey patch Panel to change the item indexing behavior.  For example, I want to monkepatch the behavior of:

     panel['item1']

Monkeypatching __getitem__ doesn't work, so I presume panel is calling an item indexer of some sort.   I've not worked closely enough with Panel to know if it would be straightforward to get this to work.  This solution is really just a current workaround to something we're doing, so we don't plan to use it forever.

Thanks

--
You received this message because you are subscribed to the Google Groups "PyData" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pydata+unsubscribe-/JYPxA39Uh5TLH3MbocFFw@public.gmane.org.
For more options, visit https://groups.google.com/d/optout.
Ryan Nelson | 2 Feb 17:28 2015
Picon

HDF memory usage question

I'm trying to extract a single column from a large HDF table, but the memory usage seems to be the same as if I select the entire HDF table. How can I limit the memory usage? Details below. (Window 7, Python 3.4, Pandas 0.15.2)

All data tables in the HDF file were created using Pandas. The largest table (right now ~100000 rows by 500 columns) was created with two columns set as data_columns: "col_1" and "col_2". I want to be able to filter sections of this large table using certain values from "col_1". I'm doing the following:

import pandas as pd
h5 = pd.HDFStore('all_data.h5')
t_df = h5.select('all_data_table', where='col_2 != 0', columns=['col_1'])

However, the memory usage here is very close to this:

import pandas as pd
h5 = pd.HDFStore('all_data.h5')
t_df = h5.select('all_data_table')

I thought that the first selection works on-disk, and would only return a single column of data, which should be a small set of data. I'm worried that when my large table gets too large for my available memory, it will cause some problems. Am I missing something in the selection process?

Thanks


--
You received this message because you are subscribed to the Google Groups "PyData" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pydata+unsubscribe-/JYPxA39Uh5TLH3MbocFFw@public.gmane.org.
For more options, visit https://groups.google.com/d/optout.
nir izraeli | 31 Jan 23:12 2015
Picon

Sparse MultiIndexed Panel not implemented?

I've been using pd.DataFrame to hold a matrix of collected features for a set of samples. 
The data arives as a stream of data points for each sample and is collected sequentially for a fixed amount of
time and each sample's vector holds an aggregated sum.
The matix is used with a few ML algorithms to classify the samples. 
For debugging purposes i want to also be abel to maintain an ordered sequence of data points. 
For that, instead of holding a sample*feature 2d matrix i thought i should have a sample*index in
sequence*feature 3d matrix, using a pd.Panel. all works well until i switch to a sparse panel, then having
the features MultiIndexed i get a NotImplementedError exception. By removing any one of: Sparse,
MultiIndex, Panel everything works but the three combined fail. 

I took a glimpse in the code and it seems there's an raise explicitly for that case. I wanted to know if it's
unsolvable or am i doing something wrong or should do it otherwise. 

Thanks!

--

-- 
You received this message because you are subscribed to the Google Groups "PyData" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pydata+unsubscribe@...
For more options, visit https://groups.google.com/d/optout.
Briti Sundar Deb | 31 Jan 08:50 2015
Picon

Pandas installation Help :)

Hi .. I have some issue in running pandas package in python 3.4 on windows.
Here is the issue when I try to import pandas it shows a strange error. It seems to be some issue with the six package which i have also installed version 1.2.0 ...Here is the code ................
####################################3
import pandas
cannot import name 'string_types'
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
File "C:\Python34\lib\site-packages\pandas\__init__.py", line 6, in <module>
from . import hashtable, tslib, lib
File "tslib.pyx", line 37, in init pandas.tslib (pandas\tslib.c:62410)
File "C:\Python34\lib\site-packages\dateutil\tz.py", line 16, in <module>
from six import string_types, PY3
ImportError: cannot import name 'string_types'

######################################

Any insight on this issue .. Thnx in advanced !

--
You received this message because you are subscribed to the Google Groups "PyData" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pydata+unsubscribe-/JYPxA39Uh5TLH3MbocFFw@public.gmane.org.
For more options, visit https://groups.google.com/d/optout.
Nick Eubank | 30 Jan 20:28 2015
Picon

Does HDF drop all NaN rows?

Hi All,


Trying to save files to HDF, and finding that when I call them back, rows where all columns were NAN seem to be missing -- is this intentional, and if so, why are we doing this? 

I understand this may feel efficient, but it's kind of unexpected, and seems potentially dangerous as a default behavior. Missing values aren't inherently meaningless -- I want to know if there are index values associated with missing values...

Nick

--
You received this message because you are subscribed to the Google Groups "PyData" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pydata+unsubscribe-/JYPxA39Uh5TLH3MbocFFw@public.gmane.org.
For more options, visit https://groups.google.com/d/optout.
Richard Styron | 30 Jan 18:12 2015
Picon

Pandas resources in Spanish

Hi,

I've got some Colombian collaborators who need to use Pandas/ IPython Notebook to explore results of our work.  I would like to send them links to tutorials or other documentation for Pandas, preferably in Spanish.  I don't see any links to Pandas docs in Spanish (or other languages for that matter) and I haven't been successful at finding anything on the internet. Does anyone know of any suitable resources?

Thanks,
Richard

--
You received this message because you are subscribed to the Google Groups "PyData" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pydata+unsubscribe-/JYPxA39Uh5TLH3MbocFFw@public.gmane.org.
For more options, visit https://groups.google.com/d/optout.
blrman91504 | 28 Jan 03:13 2015
Picon

pandsa, anaconda, numpy

i d/l pandas,and installed it.
in my script i typed 'import pandas as pd'.
when i finished typing my script, i tried to run it. i got 'syntax error line 1'.
i made sure the first character of the import statement was against the left hand side of the editor[pyscripter].

i also d/l ananconda and installed it. when i tried to run my script with reference to anaconda, i got the following error:'no module named anaconda'.
so, i  re-d/l anaconda. my system said 'anaconda already installed'.

so, i d/l numpy. i installed it too. when i tired to run my script, i got the same msg:' no module named numpy'.
so,what gives? im new to python and those packages.

blrman91504-/E1597aS9LQAvxtiuMwx3w@public.gmane.org

.

--
You received this message because you are subscribed to the Google Groups "PyData" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pydata+unsubscribe-/JYPxA39Uh5TLH3MbocFFw@public.gmane.org.
For more options, visit https://groups.google.com/d/optout.
Yuri D'Elia | 28 Jan 00:26 2015

read_table: column names in header "comment"

I have many files of the form:

  #COL1	COL2
  1	2
  ...	...

where '#' is not really used as a comment, but as a quick check for the
presence of column names.

I'm currently reading these files with read_table() and just
checking/renaming the first column, since it's less cumbersome than
opening the file, reading the first byte, and then using read_table on
the rest (although you could abort earlier in certain cases).

I'm not sure on how common is this format, but I've seen it often enough
in bioinformatics to start using it as well, since it plays well in a
number of corner cases.

Is there a better way to read this format?

--

-- 
You received this message because you are subscribed to the Google Groups "PyData" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pydata+unsubscribe@...
For more options, visit https://groups.google.com/d/optout.

Tansu Demirbilek | 25 Jan 14:39 2015
Picon

Copula analysis

Are there any plans to add copula estimation and random number generation from a copula to statsmodels or pandas? 

--
You received this message because you are subscribed to the Google Groups "PyData" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pydata+unsubscribe-/JYPxA39Uh5TLH3MbocFFw@public.gmane.org.
For more options, visit https://groups.google.com/d/optout.
AN6U5 | 25 Jan 04:12 2015
Picon

STL Decomposition

Hi All,

I'm curious if there are plans in the works to add STL decomposition to any of the Python based stats/data science packages, or if there is any capability in this arena that I've missed?  I'm not seeing anything when I search Google, Stackoverflow, this message board, or Github.  It sure would be a great addition to the time series forecasting capabilities of Pandas.

Also, if anyone has good examples of running R stl from within python, I'd love to see it.

Thanks,

Angus

--
You received this message because you are subscribed to the Google Groups "PyData" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pydata+unsubscribe-/JYPxA39Uh5TLH3MbocFFw@public.gmane.org.
For more options, visit https://groups.google.com/d/optout.

Gmane