如何在Pandas中从UTC时间序列获取星期几？遇类型错误求助

Hey there! I see the issue you're running into—let's break this down and fix it quickly.

Why the Error Happens #

The datetime.fromtimestamp() function only works with single numeric values, but you're passing an entire Pandas Series (df['utc_time']/1000) to it. Pandas can't convert the whole Series to a single int, hence the TypeError: cannot convert the series to <class 'int'>.

Solution 1: Use Pandas' Built-in Vectorized Operations (Recommended) #

Pandas has optimized methods for handling datetime conversions on entire Series—this is way faster than looping through each element, especially with large datasets.

import pandas as pd

# Convert millisecond timestamp to datetime object (vectorized operation)
df['utc_time'] = pd.to_datetime(df['utc_time'], unit='ms')

# Extract full weekday name (e.g., "Monday", "Tuesday")
df['Week_Day'] = df['utc_time'].dt.day_name()

• unit='ms' tells Pandas your timestamp is in milliseconds (your 13-digit value is a clear sign of this).
• dt.day_name() returns the full weekday name, exactly like strftime("%A") does. For older Pandas versions (pre-1.1.0), use dt.weekday_name instead.

Solution 2: Use apply() for Element-wise Processing #

If you specifically want to use the datetime module, you can use apply() to run the conversion on each individual value in the Series:

from datetime import datetime

# Apply the conversion to each element in the Series
df['Week_Day'] = df['utc_time'].apply(lambda x: datetime.fromtimestamp(x/1000).strftime("%A"))

• apply() iterates over every value in df['utc_time'], passes each single int value x to the lambda function, and runs the timestamp conversion on it. This avoids the TypeError because we're no longer passing a Series to datetime.fromtimestamp().

Quick Test to Verify #

Let's confirm with your sample timestamp:

import pandas as pd
from datetime import datetime

# Sample data with your timestamp
test_df = pd.DataFrame({'utc_time': [1536444323321]})

# Method 1
test_df['datetime'] = pd.to_datetime(test_df['utc_time'], unit='ms')
test_df['weekday_1'] = test_df['datetime'].dt.day_name()

# Method 2
test_df['weekday_2'] = test_df['utc_time'].apply(lambda x: datetime.fromtimestamp(x/1000).strftime("%A"))

print(test_df)

Both methods will return Monday (your sample timestamp converts to 2018-09-09 UTC, which is a Monday).

内容的提问来源于stack exchange，提问作者John