使用.loc选取Pandas中DatetimeIndex范围内的行（Python3）

Got it, let's get this sorted for you! When working with a pandas Series that has a DatetimeIndex, using .loc[] to grab rows within a specific date range is straightforward—you just need to nail a few key steps to make sure you get the expected subset as a DataFrame.

Step 1: Verify Your DatetimeIndex Type #

First, double-check that your Series index is actually a DatetimeIndex (not strings or another type). If it's not, pandas won't interpret your date slice correctly.

To check:

print(series.index.dtype)

If it returns something like object (meaning strings), convert it to datetime:

series.index = pd.to_datetime(series.index)

Step 2: Slice the Date Range with .loc[] #

Pandas supports inclusive range slicing with string dates when you have a DatetimeIndex. This means both the start and end dates you specify will be included in the result.

Use this syntax to get your desired range:

# Slice the Series to get rows between 2010-11-01 and 2010-12-30 (inclusive)
subset_series = series.loc['2010-11-01':'2010-12-30']

Step 3: Convert to DataFrame #

Since you want a DataFrame instead of a Series, use the .to_frame() method to convert the sliced subset:

# Convert the sliced Series to a DataFrame
subset_df = subset_series.to_frame()

# Or do it in one line
subset_df = series.loc['2010-11-01':'2010-12-30'].to_frame()

Troubleshooting Common Issues #

If you weren't getting the expected results before, here are a few things to check:

• String vs. Datetime Index: As mentioned earlier, string indexes won't slice correctly by date order—always ensure your index is a DatetimeIndex.
• Timezone Mismatches: If your index has a timezone attached but your slice dates don't, you'll run into errors. Either remove the timezone from the index:

  series.index = series.index.tz_localize(None)
  

  Or attach the same timezone to your slice timestamps:

  from pytz import UTC
  subset_series = series.loc[pd.Timestamp('2010-11-01', tz=UTC):pd.Timestamp('2010-12-30', tz=UTC)]
  

• Date Format Inconsistencies: Make sure your slice dates match the format of your index (though pandas is usually flexible with this, e.g., YYYY-MM-DD vs YYYY/MM/DD).

Full Example #

Here's a complete working example to test with:

import pandas as pd
import numpy as np

# Create a sample Series with DatetimeIndex
dates = pd.date_range(start='2010-10-20', end='2011-01-10', freq='D')
sample_series = pd.Series(np.random.randn(len(dates)), index=dates)

# Get the desired date range and convert to DataFrame
result_df = sample_series.loc['2010-11-01':'2010-12-30'].to_frame()

# Check the first and last rows to confirm the range
print("First 5 rows:")
print(result_df.head())
print("\nLast 5 rows:")
print(result_df.tail())

内容的提问来源于stack exchange，提问作者anon