BUG: HDFStore.select ignores start and stop parameters

[TomAugspurger]

Or I'm doing this wrong

In [1]: df = pd.DataFrame(np.random.randn(10, 4))

In [3]: df.to_hdf('test.h5', 'df')

In [4]: with pd.get_store('test.h5') as store:
   ...:     df_ = store.select('df', start=0, stop=2)
   ...:     

In [5]: df_
Out[5]: 
          0         1         2         3
0  1.504814  0.209394 -0.360703 -0.339684
1  0.272718  0.768927 -1.061029  0.919742
2  0.315669  0.337206 -1.358443  1.223715
3 -0.414095  0.015512 -0.320634  0.703489
4  0.889063 -0.194524  1.342604  1.078340
5 -1.636940 -0.725176 -0.096803  0.346356
6 -1.359400 -0.588007  0.651720 -0.127604
7 -1.613468  1.131560 -0.074385 -0.830540
8  0.176463 -1.298918 -0.215537  0.984911
9 -0.372920  0.898528 -0.869616  1.178704

Docstring says:

start : integer (defaults to None), row number to start selection
stop  : integer (defaults to None), row number to stop selection

I can look tomorrow (catching up on pandas stuff all day tomorrow hopefully)

[jreback]

that's a fixed store

only table implements these

they could be for a fixed store (though inefficient I think)

[TomAugspurger]

OK thanks. I'll add a note to the docstring.

[jreback]

actually it should work - maybe raise NotImplemented error for now

http://pytables.github.io/usersguide/libref/homogenous_storage.html#the-array-class

[jreback]

here's partway to an implementation of this: http://stackoverflow.com/questions/29069592/read-n-rows-from-an-h5-file/29089216#29089216