如何将datetime64[D]类型的date_index转换为月份名称？

Hey there! You’ve already got your datetime64[D] date sequence set up, so let’s break down how to convert it to both month numbers (1-12) and month names. I’ll cover pure NumPy methods and more intuitive Pandas approaches—pick what fits your workflow!

1. Convert to Month Numbers (1-12) #

Option 1: Pure NumPy Implementation #

You can leverage NumPy’s datetime64 type directly to extract month values:

import numpy as np

# Your existing date sequence
date_index = np.arange('2015-01-01','2016-01-01', dtype='datetime64[D]')

# Convert to month numbers (1 to 12)
month_numbers = (date_index.astype('datetime64[M]').astype(int) % 12) + 1

Here’s the breakdown: We first cast the dates to month precision (datetime64[M]), convert those to integers, take modulo 12 to wrap around the year, then add 1 to shift from 0-11 to 1-12.

Option 2: Pandas (More Readable) #

If you’re already using Pandas in your project, this method is cleaner and easier to follow:

import numpy as np
import pandas as pd

date_index = np.arange('2015-01-01','2016-01-01', dtype='datetime64[D]')
# Convert to Pandas DatetimeIndex
pd_dates = pd.DatetimeIndex(date_index)
# Grab month numbers directly
month_numbers = pd_dates.month

2. Convert to Month Names (e.g., "January", "Feb") #

Pandas really shines here with built-in methods for month names, but I’ll also include a pure NumPy option if you want to avoid dependencies.

Option 1: Pandas for English Month Names #

Get full names or abbreviations with just a couple lines:

import numpy as np
import pandas as pd

date_index = np.arange('2015-01-01','2016-01-01', dtype='datetime64[D]')
pd_dates = pd.DatetimeIndex(date_index)

# Full month names (e.g., "January", "February")
full_month_names = pd_dates.month_name()

# Shortened month names (e.g., "Jan", "Feb")
short_month_names = pd_dates.month_abbr

Bonus: If you need month names in another language, you can set the locale (e.g., pd_dates.month_name(locale='fr_FR') for French).

Option 2: Pure NumPy (Custom Mapping) #

If you don’t want to use Pandas, create a month name mapping and match it to your month numbers:

import numpy as np

date_index = np.arange('2015-01-01','2016-01-01', dtype='datetime64[D]')
# First get month numbers using the NumPy method above
month_numbers = (date_index.astype('datetime64[M]').astype(int) % 12) + 1

# Define your month name list (index 0 is unused to match 1-12)
month_name_map = [
    "", "January", "February", "March", "April", "May", "June",
    "July", "August", "September", "October", "November", "December"
]

# Map numbers to names
month_names = np.array(month_name_map)[month_numbers]

内容的提问来源于stack exchange，提问作者Razieh Delavar