Updating dataframe in R without for loop

Question

I have one master dataframe df:

df <- data.frame(c("A", "B", "C"), c(1,2,3), c(3,1,2), c(4,2,1), rep(NA, 3), rep(NA, 3))
colnames(df) <- c("text", "var1", "var2", "var3", "value1", "value2")

And another dataframe df.upd with new information:

df.upd <- data.frame(c(1,2), c(3,1), c(4,2),c(0.5, 0.6), c(12, 20))                           
colnames(df.upd) <- c("var1", "var2", "var3", "value1", "value2")
> df
text var1 var2 var3 value1 value2
1    A    1    3    4     NA     NA
2    B    2    1    2     NA     NA
3    C    3    2    1     NA     NA
> df.upd
  var1 var2 var3 value1 value2
1    1    3    4    0.5     12
2    2    1    2    0.6     20

I want to match columns "var1", "var2", "var3" and update the columns "value1" and "value2". So row 1 and 2 of df.upd would update row 1 and 2 of df, ergo as.numeric(df.upd[row x, 1:3])==as.numeric(df[row y, 2:4]) must be TRUE.

The master df has around 30k rows and 60 columns, so a for loop is not an option. Any idea how to accomplish this faster?

you can use merge with all.x = TRUE (left join), then use ifelse with is.na to update the relevant columns, then drop extra columns — moodymudskipper
– moodymudskipper, Commented May 23, 2018 at 8:20
check out stackoverflow.com/questions/36347213/…... i.e. library(data.table); cols <- paste0('value', 1:2); setDT(df)[setDT(df.upd), (cols) := mget(paste0("i.", cols)), on=.(var1, var2, var3)] — chinsoon12
– chinsoon12, Commented May 23, 2018 at 8:25

SeGa · Accepted Answer · 2018-05-23 08:42:57Z

0

It's a for-loop answer, but it still might be useful and fast, as the process is vectorized.

ind <- intersect(names(df), names(df.upd))
for (i in ind) {
  df[1:length(df.upd[,i]),i]  <- df.upd[,i]
}

answered May 23, 2018 at 8:42

SeGa

9,8793 gold badges34 silver badges76 bronze badges

Sign up to request clarification or add additional context in comments.

Collectives™ on Stack Overflow

Updating dataframe in R without for loop

1 Answer 1

Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related