My data.table has many columns in groups of four. On each group, I want
to perform the same set of calculations.
mydt <- data.table(
date = seq(as.IDate("2025-06-01"), by = 1, length.out = 10),
ABC1_price = runif(10, 100, 120), ABC1_volume = runif(10, 200, 300),
ABC2_price = runif(10, 100, 120), ABC2_volume = runif(10, 200, 300),
DEF1_price = runif(10, 100, 120), DEF1_volume = runif(10, 200, 300),
DEF2_price = runif(10, 100, 120), DEF2_volume = runif(10, 200, 300))
mydt[
, let(ABC_price = fifelse(ABC1_volume < ABC2_volume, ABC2_price,
ABC1_price),
DEF_price = fifelse(DEF1_volume < DEF2_volume, DEF2_price, DEF1_price))
]
# Now use mydt[, list(ABC_price, DEF_price)] for subsequent calculations
With many columns, above method is error-prone and tedious. Is it
possible to do something like:
myvars <- c("ABC", "DEF")
for (avar in myvars) {
# code for above calculations
}
Thanks,
Naresh
If I understand correctly, see ?"[" : data tables inherit from data
frames,
and data frame columns can be indexed by their names as character strings.
The following should give you the idea:
d <- data.frame(a1 = 1:3, a2 = 4:6)
x <- vector("numeric",2)
for(i in 1:2){
nm <- paste0("a",i)
x[i] <- sum(d[, nm])
}
x
Please note: This is **absoutely** awful R code just meant to illustrate
the sort of thing that you want to do, not as a strict model to be followed.
Also, as I don't use data tables, there may be better ways to do this using
data table's enhanced functionality directly.
Cheers,
Bert
On Wed, Jun 18, 2025 at 3:27?AM Naresh Gurbuxani <
naresh_gurbuxani at hotmail.com> wrote:
> My data.table has many columns in groups of four. On each group, I want
> to perform the same set of calculations.
>
> mydt <- data.table(
> date = seq(as.IDate("2025-06-01"), by = 1, length.out = 10),
> ABC1_price = runif(10, 100, 120), ABC1_volume = runif(10, 200, 300),
> ABC2_price = runif(10, 100, 120), ABC2_volume = runif(10, 200, 300),
> DEF1_price = runif(10, 100, 120), DEF1_volume = runif(10, 200, 300),
> DEF2_price = runif(10, 100, 120), DEF2_volume = runif(10, 200, 300))
>
> mydt[
> , let(ABC_price = fifelse(ABC1_volume < ABC2_volume, ABC2_price,
> ABC1_price),
> DEF_price = fifelse(DEF1_volume < DEF2_volume, DEF2_price, DEF1_price))
> ]
>
> # Now use mydt[, list(ABC_price, DEF_price)] for subsequent calculations
>
> With many columns, above method is error-prone and tedious. Is it
> possible to do something like:
>
> myvars <- c("ABC", "DEF")
> for (avar in myvars) {
> # code for above calculations
> }
>
> Thanks,
> Naresh
>
> ______________________________________________
> R-help at r-project.org mailing list -- To UNSUBSCRIBE and more, see
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide
> https://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
>
[[alternative HTML version deleted]]
Here is a "brute strength and stupidity" approach that may help.
Note I renamed your data.table to DT for my own convenience.
##===========================================##
# Load packages ------------------------------------------------------------
suppressMessages(library(data.table))
# Load dada. --------------------------------------------------------------
DT <- fread("raw.data.csv"); DT
# Subset data and rbind into one new data.table. -------------------------
mynames <- c("date", "price1", "vol1",
"price2","vol2", "id")
ABC <- DT[, c(1, 2:5)][, id := "ABC"] ; names(ABC) <- mynames
DEF <- DT[, c(1, 6:9)][, id := "DEF"] ; names(DEF) <- mynames
DT1 <- rbind(ABC, DEF)
# Create new variable "price".
--------------------------------------------
DT1[, price := fcase(vol1 < vol2, price2, default = price1)][]
# Summary statistic using "by".
-------------------------------------------
DT1[, .(mysum = sum(price)), by = id]
##===========================================##
On Wed, 18 Jun 2025 at 06:27, Naresh Gurbuxani <naresh_gurbuxani at
hotmail.com>
wrote:
> My data.table has many columns in groups of four. On each group, I want
> to perform the same set of calculations.
>
> mydt <- data.table(
> date = seq(as.IDate("2025-06-01"), by = 1, length.out = 10),
> ABC1_price = runif(10, 100, 120), ABC1_volume = runif(10, 200, 300),
> ABC2_price = runif(10, 100, 120), ABC2_volume = runif(10, 200, 300),
> DEF1_price = runif(10, 100, 120), DEF1_volume = runif(10, 200, 300),
> DEF2_price = runif(10, 100, 120), DEF2_volume = runif(10, 200, 300))
>
> mydt[
> , let(ABC_price = fifelse(ABC1_volume < ABC2_volume, ABC2_price,
> ABC1_price),
> DEF_price = fifelse(DEF1_volume < DEF2_volume, DEF2_price, DEF1_price))
> ]
>
> # Now use mydt[, list(ABC_price, DEF_price)] for subsequent calculations
>
> With many columns, above method is error-prone and tedious. Is it
> possible to do something like:
>
> myvars <- c("ABC", "DEF")
> for (avar in myvars) {
> # code for above calculations
> }
>
> Thanks,
> Naresh
>
> ______________________________________________
> R-help at r-project.org mailing list -- To UNSUBSCRIBE and more, see
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide
> https://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
>
--
John Kane
Kingston ON Canada
[[alternative HTML version deleted]]
? Wed, 18 Jun 2025 10:27:37 +0000 Naresh Gurbuxani <naresh_gurbuxani at hotmail.com> ?????:> mydt <- data.table( > date = seq(as.IDate("2025-06-01"), by = 1, length.out = 10), > ABC1_price = runif(10, 100, 120), ABC1_volume = runif(10, 200, 300), > ABC2_price = runif(10, 100, 120), ABC2_volume = runif(10, 200, 300), > DEF1_price = runif(10, 100, 120), DEF1_volume = runif(10, 200, 300), > DEF2_price = runif(10, 100, 120), DEF2_volume = runif(10, 200, 300)) > > mydt[ > , let(ABC_price = fifelse(ABC1_volume < ABC2_volume, ABC2_price, > ABC1_price), > DEF_price = fifelse(DEF1_volume < DEF2_volume, DEF2_price, > DEF1_price)) ]Thank you for the reproducible example! How about the following? for (var in c("ABC", "DEF")) mydt[, price := fifelse(volume1 < volume2, price2, price1), env = list( price = paste0(var, '_price'), price1 = paste0(var, '1_price'), price2 = paste0(var, '2_price'), volume1 = paste0(var, '1_volume'), volume2 = paste0(var, '2_volume') ) ] -- Best regards, Ivan