Difference between parse_date=[0] and parse_date=True in pandas.read_csv

pandas

Solution

Looks like a bug to me. I created an issue for this.

Note that by using the *index_col* argument it is possible to set the index.

In [15]: df = pd.read_csv(StringIO(data),parse_dates=[0], index_col=0)

In [15]: df.index
<class 'pandas.tseries.index.DatetimeIndex'>
[2012-07-31 02:00:00, ..., 2012-07-31 02:30:00]
Length: 3, Freq: None, Timezone: None

Problem

This code: ``` import pandas as pd from StringIO import StringIO data = "date,c1\n2012-07-31 02:00,1.1\n2012-07-31 02:15,2.2\n2012-07-31 02:30,3.3\n" df1 = pd.read_csv(StringIO(data),parse_dates=True).set_index(('date')) df2 = pd.read_csv(StringIO(data),parse_dates=[0] ).set_index(('date')) print "df1:\n{index}".format(index=df1.index) print "df2:\n{index}".format(index=df2.index) ``` returns: ``` df1: array([2012-07-31 02:00, 2012-07-31 02:15, 2012-07-31 02:30], dtype=object) df2: <class 'pandas.tseries.index.DatetimeIndex'> [2012-07-31 02:00:00, ..., 2012-07-31 02:30:00] Length: 3, Freq: None, Timezone: None ``` Is this difference between df1 and df2 a bug,feature, or have I misunderstood something?

Original source