如何用data.table的setcolorder按另一数据框列重排数据框列

Hey there! Let's sort out this error you're seeing with setcolorder. The problem here is pretty straightforward: the setcolorder function doesn't accept a key= argument — that's not one of its parameters.

Why You Got the Error #

When you ran gf3<-setcolorder(df1, key=df2$names), you tried to pass the column order vector using a named argument that doesn't exist. setcolorder expects the new column order as its second positional argument (no need for a parameter name like key).

The Correct Implementation #

First, a quick note about data.table behavior: most set* functions like setcolorder modify the data.table in-place (they change the original object directly). If you want to keep your original df1 intact, make sure to copy it first with copy().

Here's the step-by-step fix:

1. Load the data.table package (if you haven't already)
2. Create a copy of df1 to avoid modifying the original data
3. Use setcolorder with just the target data.table and your desired column order vector

library(data.table)

# Example data matching your scenario
df1 <- data.table(five = 1:5, two = 6:10, three = 11:15, one = 16:20, four = 21:25)
df2 <- data.table(names = c("one", "two", "three", "four", "five"))

# Make a copy of df1 first (critical if you want to preserve the original)
gf3 <- copy(df1)

# Reorder columns directly with df2$names — no key= argument!
setcolorder(gf3, df2$names)

# Check the result
gf3

Key Notes to Remember #

• Always use copy() if you don't want to alter the original data.table — skipping this will change df1's column order permanently.
• Double-check that every name in df2$names exists in df1 — if there's a mismatch, setcolorder will throw an error about missing columns.
• setcolorder is designed for speed (it's optimized for data.table), so it doesn't return the modified table — that's why we modify the copied table directly.

内容的提问来源于stack exchange，提问作者Jakhongir Alidjanov