R: split dataframe into vectors (one per column) and keep name of column

Asked Apr 09 '19 at 16:50

Active Apr 09 '19 at 16:53

Viewed 43 times

I have a dataframe as such:

sample gender disease treatment
sample1      F     c   starved
sample2      F     d       fed
sample3      M     d   starved
sample4      M     d       fed
sample5      M     c   starved
sample6      M     d       fed

I would like to split the dataframe to obtain one vector per colum with each vector keeping its original column name. This should be a general command (I won't always be applying to this general df).

I know to obtain a simple vector with something like:

gender <- sample$gender

But the file I will need to work with will differ between different experiments and will have different column names and numbers.

I looked at this link but it does not allow you to keep the column names (i.e. the accepted solution forced you to name your vectors Row1, Row2 etc). Any help would be deeply appreciated.

How to split a R data frame into vectors (unbind)

edited Apr 09 '19 at 16:53

asked Apr 09 '19 at 16:50

mf94

Do you need `split(df1, df1$gender)` – akrun Apr 09 '19 at 16:51
No, the names of the columns will differ between my experiment. In my example I have a column called gender but I may not in my next experiment. – mf94 Apr 09 '19 at 16:53
3

Like this? e.g., `list2env(mtcars, envir = .GlobalEnv)` – Andrew Apr 09 '19 at 16:54
4

`list2env` does this, also `attach`, but beware. There's rarely a good reason to do this, mostly it causes hard-to-find bugs when your now-separate vectors get out of sync with each other. – Gregor Thomas Apr 09 '19 at 16:56
@Andrew That's it! I was confused by the solution suggested in the link I pasted but your command is simpler and it works! – mf94 Apr 09 '19 at 16:58
1

Glad it helped, and keep @Gregor's qualifier in mind :) – Andrew Apr 09 '19 at 17:05
wow, this was something new. I never thought, we could do this. But as per the manual, this is not efficient with large datasets. – Slayer Apr 09 '19 at 17:20
1

@Prerit I mean, it's not a good idea for several reasons. It seems like an especially bad idea for large data sets, not just because of inefficiency. – Gregor Thomas Apr 09 '19 at 18:40
@Gregor I would rather stick to using `$` or use `pull` function from `dplyr`. – Slayer Apr 09 '19 at 18:42
@Prerit `fortunes::fortune(379)` – Gregor Thomas Apr 09 '19 at 18:44
@Gregor loved it. haha! – Slayer Apr 09 '19 at 18:50

R: split dataframe into vectors (one per column) and keep name of column

0 Answers0