Pandas数据处理求助：UPOR为0/NaN时用NPOR计算Qnd列

Hey there! Let's sort out your Qnd column calculation. The goal is to use UPOR when it's not 0 or NaN, and fall back to NPOR whenever those invalid cases pop up. Here are a few straightforward, efficient ways to do this:

Method 1: Use numpy.where (Intuitive Conditional Check) #

This is great for explicit, easy-to-read conditional logic:

import numpy as np

# Check if UPOR is 0 or NaN; use NPOR in those cases, else UPOR
GU_ES3['Qnd'] = GU_ES3['Qnd,hw_m2'] * np.where(
    (GU_ES3['UPOR'] == 0) | GU_ES3['UPOR'].isna(),
    GU_ES3['NPOR'],
    GU_ES3['UPOR']
)

The np.where function evaluates the condition for every row: if the condition is True (UPOR is 0 or missing), it uses the NPOR value; otherwise, it sticks with UPOR.

Method 2: Pandas where (Cleaner Pandas-Native Approach) #

Pandas has its own where method (note: it inverts the condition compared to numpy.where):

# Define when UPOR is valid (not NaN and not 0)
valid_upor = GU_ES3['UPOR'].notna() & (GU_ES3['UPOR'] != 0)

# Keep UPOR where valid, replace with NPOR otherwise
GU_ES3['Qnd'] = GU_ES3['Qnd,hw_m2'] * GU_ES3['UPOR'].where(valid_upor, GU_ES3['NPOR'])

This approach keeps your code focused on the "valid" case first, which can make logic easier to follow for complex conditions.

Method 3: replace + fillna (Concise One-Liner) #

If you prefer a more compact solution, convert 0s to NaNs first, then fill all NaNs with NPOR:

import numpy as np

GU_ES3['Qnd'] = GU_ES3['Qnd,hw_m2'] * GU_ES3['UPOR'].replace(0, np.nan).fillna(GU_ES3['NPOR'])

This works because replace(0, np.nan) turns all 0 values into missing values, then fillna(GU_ES3['NPOR']) replaces every NaN (original or converted) with the corresponding NPOR value.

All three methods will give you the same result—pick the one that fits your coding style best!

内容的提问来源于stack exchange，提问作者energyMax