.loc function failed

[goog]

env: windows 7 64bit, pandas 0.12

In [106]: res
Out[106]:
<class 'pandas.core.frame.DataFrame'>
Int64Index: 93705 entries, 13 to 93713
Columns: 3561 entries, 2) to 3561)
dtypes: float64(3561)


In [107]: res.loc[12]
Out[107]:
0
2)     3.362026
3)     1.574658
4)     1.905283
6)     2.069734
8)     4.777519
9)     4.316618
11)    2.231455
......

my res's index starts with 13, but the numeric index loc[12] has values.

[hayd]

You can check that with is_monotonic (the 13 to 93713 isn't the min and max, rather it's the first and last)

You can check whether they are in order:

df.index.is_monotonic

And check whether 12 is in the index (if it isn't, I think you've found a bug):

12 in df.index

Just to show simple example:

In [11]: df = pd.DataFrame([[1,2], [3, 4]], index=[1, 2])

In [12]: pd.concat([df]*100)[:-1]  # glue together 100 dfs but don't include last row of the final one
Out[12]:
<class 'pandas.core.frame.DataFrame'>
Int64Index: 199 entries, 1 to 1
Data columns (total 2 columns):
0    199  non-null values
1    199  non-null values
dtypes: int64(2)

In [13]: pd.concat([df]*100)[:-1].index.is_monotonic
Out[13]: False

In [14]: 2 in pd.concat([df]*100)[:-1].index
Out[14]: True

[goog]

yeah, ur right! I misunderstood the "Int64Index: 93705 entries, 13 to 93713", thanks very much!