Skip to content

Removing funny characters from a column of a data frame

2 messages · Bansal, Vikas, Joshua Wiley

Original
Bansal, Vikas
# 
Dear all,

The 5th column of my data frame is like this-

.$.$.$.$.$,$,$...,,,,,.,,.,,...,,,,.,,....,,,T...,,,,,,,,,,,.,,,,,....,,...,,
,..,,....,,,,,...,,,..,,......,,,,,,,....,,,.,,,,....,,...G.,,,,,,,,...,,,,,,.,,
,t.,,c,,.a.,,,.A,,,,....,,,.....,,,,..........,,,,,..,,,.,,,....,,,,,...,,,$....
.,,,,..,,,...,,,,,..,,,,,,.............$..,,,,,,...,,..,,$,...,,,,,,,....,,,,,,.
,,,,......,,,,.,,.......,.....,,,,,,.,,..,,...,,,,,.,......,.......,,....,,,,..,,
,,,,.........,,,,,.....,,,,...,,,.....,,.....,,......,....,,......,.,,..,,,,...,,
H.,,,..,,.....,,,,..,,,,,,,,,^~.^~.^\".^~.^~.^~.^~,^~,^~,^~,"  

I just want to have A,a,C,c,G,g,T,t and dot and comma in the columns.

example of first row should be-

.....,,...,,,,,.,,.,,...,,,,.,,....,,,T...,,,,,,,,,,,.,,,,,....,,...,,


currently i am using this code-

df$V5 <-  apply(df, 1, function(x) gsub("\\:|\\$|\\^|!|\\-|1|2|3|4|5|6|7|8|10|~|H", "",x[5]))

this use of gsub looks odd to me,although result is coming good but I want something fast because data is large.I want something like this-

delete everything else except  A,a,C,c,G,g,T,t and dot and comma.

Any suggestions Please.


        
Thanking you,
Warm Regards
Vikas Bansal
Msc Bioinformatics
Kings College London
Joshua Wiley
# 
Hi Vikas,

You're overworking yourself here, gsub is vectorized!

df$V5 <- gsub("[^AaCcGgTt\\.,]", "", df$V5)

This will be *substantially* faster than looping (using apply) over
every row of your data frame, since you just care about the 5th column
anyways.  Also, I switched your regexp for one that replaces not
AaCcGgTt.,

Cheers,

Josh
On Sun, Aug 7, 2011 at 12:57 PM, Bansal, Vikas <vikas.bansal at kcl.ac.uk> wrote: