.I behavior in by

[maxmoro]

This code

library(data.table)
t1=data.table(a=c(1,2,3),b=c(7,8,9))
t1[,':='(c=list(c(a,b))),by=t1[,.I]]
t1

Has the output I need. Basically column C is a list with vectors created from A and B within same Row

   a b   c
1: 1 7 1,7
2: 2 8 2,8
3: 3 9 3,9

But, if I write it in this way that is more clear to read:

library(data.table)
t1=data.table(a=c(1,2,3),b=c(7,8,9))
t1[,':='(c=list(c(a,b))),by=.I]
t1

It doesn't work:

   a b           c
1: 1 7 1,2,3,7,8,9
2: 2 8 1,2,3,7,8,9
3: 3 9 1,2,3,7,8,9

Why?
.I should help to group by Row number, but it is not.
Is this a bug or is intended to work in this way?

Thanks
Max

[franknarf1]

In github posts, you can write blocks of codes (over multiple lines) by indenting each line four spaces. No need for backticks when doing that.

Currently, .I does not work in by; maybe it should throw an error there? I usually do by=seq(nrow(DT)), though it's not ideal.

[maxmoro]

Thank you for the block code tip :)
No error is returned, it just works like there is no by
I like the suggestion you are offering, I'll update my code, so is more readable.

Anyway, it would be great to have the by=.I implemented!
Thanks
Max

[arunsrinivasan]

Thanks for following up, @franknarf1

[MichaelChirico]

@franknarf1 I find whole section blocking easier on GitHub (precede and follow code blocks with a line of three backticks:

```
#some code without extra indentation
```

[jangorecki]

Looks like a duplicate of #1004 please re-open if not