JJFord3 JJFord3 - 1 month ago 10
R Question

How to order a data frame using a string vector with the names of the columns to be ordered?

Lets assume the data looks like this:

df3 = data.frame(matrix(rnorm(20), nrow=10))
df3$ACCT_ID <- as.numeric(rownames(df3))
df3$X3 <- ifelse(df3$X2 < 1,"LOW","HIGH")

df3
X1 X2 ACCT_ID X3
1 0.8533321 0.73676965 1 LOW
2 -0.2289266 -0.87010747 2 LOW
3 -0.7997897 0.49979417 3 LOW
4 0.5281416 -0.70161594 4 LOW
5 -0.6246283 0.81879826 5 LOW
6 0.9939673 -1.20910819 6 LOW
7 -0.8955201 0.42167923 7 HIGH
8 1.4404649 0.59220989 8 LOW
9 -0.8687570 0.09393659 9 LOW
10 -0.7030129 -0.95184303 10 LOW


I'm trying to re-order based on X1 and ACCT_ID.

TEMP_X_DF <- df3[with(df3, order(X1,ACCT_ID)), ]
TEMP_X_DF
X1 X2 ACCT_ID X3
7 -0.8955201 0.42167923 7 HIGH
9 -0.8687570 0.09393659 9 LOW
3 -0.7997897 0.49979417 3 LOW
10 -0.7030129 -0.95184303 10 LOW
5 -0.6246283 0.81879826 5 LOW
2 -0.2289266 -0.87010747 2 LOW
4 0.5281416 -0.70161594 4 LOW
1 0.8533321 0.73676965 1 LOW
6 0.9939673 -1.20910819 6 LOW
8 1.4404649 0.59220989 8 LOW


But when I try to do this when populating from a list I only get the first row read in returned:

IDS <- c("X1","ACCT_ID")
for (id in 1:length(IDS)){
if (id == 1){
ID_OUT <- IDS[id]
}
else {
ID_OUT <- paste(ID_OUT,IDS[id],sep=",")
}
}
TEMP_X_DF <- df3[with(df3, order(ID_OUT)), ]
TEMP_X_DF
X1 X2 ACCT_ID X3
1 0.8533321 0.7367696 1 LOW


Is the string ID_OUT not resolving properly?

Answer

You could do this using get:

df3[with(df3, order(get(IDS))), ]

You should not concatenate the column names as you did in ID_OUT. Instead, you can put the desired column names in a vector and then apply get.