R help - Aug 2011 - conditional data replace (recode, change or whatsoever)

Hello,
I have a big data.frame, a piece of it as follows.

 	a	        b		        c	    d
1	58009	2010-11-02	0	    NA
2	114761	NA		        1	    2008-11-05
3	184440	NA		        1	    2009-12-08
4	189372	NA		        0	    NA
5	105286	NA		        0	    NA
6	186717	NA		        0	    NA
7	189106	NA		        0	    NA
8	127306	NA		        0	    NA
9	157342	2011-04-25	0	    NA

I want to replace b[NA] values with "20011-07-28" where c==0. I use
rstudio
and i'm a novice.


--
View this message in context:
http://r.789695.n4.nabble.com/conditional-data-replace-recode-change-or-whatsoever-tp3714715p3714715.html
Sent from the R help mailing list archive at Nabble.com.

Please do your homework before asking the list:
An introduction to R, chapter 7

Am 03.08.2011 10:05, schrieb zcatav:
> Hello,
> I have a big data.frame, a piece of it as follows.
> 
>  	a	        b		        c	    d
> 1	58009	2010-11-02	0	    NA
> 2	114761	NA		        1	    2008-11-05
> 3	184440	NA		        1	    2009-12-08
> 4	189372	NA		        0	    NA
> 5	105286	NA		        0	    NA
> 6	186717	NA		        0	    NA
> 7	189106	NA		        0	    NA
> 8	127306	NA		        0	    NA
> 9	157342	2011-04-25	0	    NA
> 
> I want to replace b[NA] values with "20011-07-28" where c==0. I
use rstudio
> and i'm a novice.
> 
> 
> --
> View this message in context:
http://r.789695.n4.nabble.com/conditional-data-replace-recode-change-or-whatsoever-tp3714715p3714715.html
> Sent from the R help mailing list archive at Nabble.com.
> 
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide
http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.

Hi
> 
> Hello,
> I have a big data.frame, a piece of it as follows.
> 
>     a           b              c       d
> 1   58009   2010-11-02   0       NA
> 2   114761   NA              1       2008-11-05
> 3   184440   NA              1       2009-12-08
> 4   189372   NA              0       NA
> 5   105286   NA              0       NA
> 6   186717   NA              0       NA
> 7   189106   NA              0       NA
> 8   127306   NA              0       NA
> 9   157342   2011-04-25   0       NA
> 
> I want to replace b[NA] values with "20011-07-28" where c==0. I
use
rstudio
> and i'm a novice.
I believe there are better solutions but I would use two steps

select rows where c==0 (see also FAQ 7.31)
sel<-which(big.data.frame$c==0)

change NA values in b column based on sel
big.data.frame$b[sel][is.na(big.data.frame$b[sel])]<-"20011-07-28"

Beware of data types AFAIK R can not accept "20011-07-28" as a date.

Regards
Petr
> 
> 
> --
> View this message in context: http://r.789695.n4.nabble.com/conditional-
> data-replace-recode-change-or-whatsoever-tp3714715p3714715.html
> Sent from the R help mailing list archive at Nabble.com.
> 
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide 
http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.

Petr Pikal wrote:
> 
> Hi
> I believe there are better solutions but I would use two steps
> 
> select rows where c==0 (see also FAQ 7.31)
> sel<-which(big.data.frame$c==0)
> 
> change NA values in b column based on sel
>
big.data.frame$b[sel][is.na(big.data.frame$b[sel])]<-"20011-07-28"
> 
> Beware of data types AFAIK R can not accept "20011-07-28" as a
date.
> 
> Regards
> Petr
> 
> 
Thanks, it runs like a charm.
Replaced date format just a typo.

--
View this message in context:
http://r.789695.n4.nabble.com/conditional-data-replace-recode-change-or-whatsoever-tp3714715p3715080.html
Sent from the R help mailing list archive at Nabble.com.

As others have noted, this is discussed in many free R tutorials, but if you
want to do it in one line I think this should do it:

X[is.NA(X[,"b"])&(X[,"c"]==0),"b"]<-"2011-07-28"
#where X is the name of the data frame.

It's a somewhat convoluted line of code but if you read it inside out the
logic is clear:

Find those rows where column b is NA and c is 1 by searching all rows of the
relevant column (the X[,something] syntax): select those rows and the b column.
Put the desired date in those slots. let me know of I can further clarify this.
I changed the date assuming a typo on your end.

Welcome and good luck getting started with R,

Michael Weylandt


On Aug 3, 2011, at 4:05 AM, zcatav <zcatav at gmail.com> wrote:
> Hello,
> I have a big data.frame, a piece of it as follows.
> 
>   a            b                c        d
> 1    58009    2010-11-02    0        NA
> 2    114761    NA                1        2008-11-05
> 3    184440    NA                1        2009-12-08
> 4    189372    NA                0        NA
> 5    105286    NA                0        NA
> 6    186717    NA                0        NA
> 7    189106    NA                0        NA
> 8    127306    NA                0        NA
> 9    157342    2011-04-25    0        NA
> 
> I want to replace b[NA] values with "20011-07-28" where c==0. I
use rstudio
> and i'm a novice.
> 
> 
> --
> View this message in context:
http://r.789695.n4.nabble.com/conditional-data-replace-recode-change-or-whatsoever-tp3714715p3714715.html
> Sent from the R help mailing list archive at Nabble.com.
> 
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide
http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.

Your suggestion works perfect as i pointed previous message. Now have another
question about data editing. I try this code:
X[X[,"c"]==1,"b"]<-X[,"d"]
and results with error: `[<-.data.frame`(`*tmp*`, X[, "c"] == 1,
"b", value
= c(NA,  : 
  replacement has 9 rows, data has 2

Logically i selected 2 rows with X[,"c"]==1. Than i want to replace in
that
rows its own data from "d" to "b" with
X[,"b"]<-X[,"d"]. What is wrong?

--
View this message in context:
http://r.789695.n4.nabble.com/conditional-data-replace-recode-change-or-whatsoever-tp3714715p3715218.html
Sent from the R help mailing list archive at Nabble.com.

On Wed, Aug 3, 2011 at 8:09 AM, zcatav <zcatav at gmail.com>
wrote:
> Your suggestion works perfect as i pointed previous message. Now have
another
> question about data editing. I try this code:
> X[X[,"c"]==1,"b"]<-X[,"d"]
> and results with error: `[<-.data.frame`(`*tmp*`, X[, "c"] ==
1, "b", value
> = c(NA, ?:
> ?replacement has 9 rows, data has 2
>
> Logically i selected 2 rows with X[,"c"]==1. Than i want to
replace in that
> rows its own data from "d" to "b" with
X[,"b"]<-X[,"d"]. What is wrong?
>
Also check out transform and ifelse, e.g.

transform(X, b = ifelse(is.na(b) & c == 0, "2011-07-28", b))

transform(X, b = ifelse(c == 1, d, c))


-- 
Statistics & Software Consulting
GKX Group, GKX Associates Inc.
tel: 1-877-GKX-GROUP
email: ggrothendieck at gmail.com

On Aug 3, 2011, at 8:09 AM, zcatav wrote:
> Your suggestion works perfect as i pointed previous message. Now  
> have another
> question about data editing. I try this code:
> X[X[,"c"]==1,"b"]<-X[,"d"]
> and results with error: `[<-.data.frame`(`*tmp*`, X[, "c"] ==
1,
> "b", value
> = c(NA,  :
>  replacement has 9 rows, data has 2
>
> Logically i selected 2 rows with X[,"c"]==1. Than i want to
replace
> in that
> rows its own data from "d" to "b" with
X[,"b"]<-X[,"d"]. What is
> wrong?
You need to apply the same logical test/selection on the rows of the  
RHS as you are doing on the LHS.
Possibly:

X[ X[,"c"]==1, "b"] <- X[ X[,"c"]==1,
"d"]

(No data, not tested code.)
-- 

David Winsemius, MD
West Hartford, CT

Gabor Grothendieck wrote:
> 
> On Wed, Aug 3, 2011 at 8:09 AM, zcatav <zcatav at gmail.com>
wrote:
>> Your suggestion works perfect as i pointed previous message. Now have
>> another
>> question about data editing. I try this code:
>> X[X[,"c"]==1,"b"]<-X[,"d"]
>> and results with error: `[<-.data.frame`(`*tmp*`, X[, "c"]
== 1, "b",
>> value
>> = c(NA, ?:
>> ?replacement has 9 rows, data has 2
>>
>> Logically i selected 2 rows with X[,"c"]==1. Than i want to
replace in
>> that
>> rows its own data from "d" to "b" with
X[,"b"]<-X[,"d"]. What is wrong?
>>
> 
> Also check out transform and ifelse, e.g.
> 
> transform(X, b = ifelse(is.na(b) & c == 0, "2011-07-28", b))
> 
> transform(X, b = ifelse(c == 1, d, c))
> 
> 
> transform(X, b = ifelse(is.na(b) & c == 0, "2011-07-28", b))
This code results as follows. Data at [1,b] and [9,b] not managed as Date.
       a          b      c          d
1  58009      14915 0       <NA>
2 114761       <NA> 1 2008-11-05
3 184440       <NA> 1 2009-12-08
4 189372 2011-07-28 0       <NA>
5 105286 2011-07-28 0       <NA>
6 186717 2011-07-28 0       <NA>
7 189106 2011-07-28 0       <NA>
8 127306 2011-07-28 0       <NA>
9 157342      15089 0       <NA>


And the second code
> transform(X, b = ifelse(c == 1, d, c))
results as follows. Data at [,b] are completly lost.

       a     b c          d
1  58009     1 0       <NA>
2 114761 14188 1 2008-11-05
3 184440 14586 1 2009-12-08
4 189372     1 0       <NA>
5 105286     1 0       <NA>
6 186717     1 0       <NA>
7 189106     1 0       <NA>
8 127306     1 0       <NA>
9 157342     1 0       <NA>

I think this solution not proper for me.

--
View this message in context:
http://r.789695.n4.nabble.com/conditional-data-replace-recode-change-or-whatsoever-tp3714715p3715525.html
Sent from the R help mailing list archive at Nabble.com.

David Winsemius wrote:
> 
> On Aug 3, 2011, at 8:09 AM, zcatav wrote:
> ........................
> You need to apply the same logical test/selection on the rows of the  
> RHS as you are doing on the LHS.
> Possibly:
> 
> X[ X[,"c"]==1, "b"] <- X[ X[,"c"]==1,
"d"]
> 
> 
This solution was suggested by R. Michael Weylandt and it works great.


--
View this message in context:
http://r.789695.n4.nabble.com/conditional-data-replace-recode-change-or-whatsoever-tp3714715p3715544.html
Sent from the R help mailing list archive at Nabble.com.

zcatav <zcatav <at> gmail.com> writes:
> 
> Your suggestion works perfect as i pointed previous message. Now have
another
> question about data editing. I try this code:
> X[X[,"c"]==1,"b"]<-X[,"d"]
> and results with error: `[<-.data.frame`(`*tmp*`, X[, "c"] ==
1, "b", value
> = c(NA,  : 
>   replacement has 9 rows, data has 2
> 
is this equivalent and/or preferred to:

X$b[X$c==1]<-X$d[X$c==1] ??

I assume this goes back to the various indexing methods for a dataframe, an
object vector that is a column of a data frame vs. an object data frame that
happens to be one column of a larger data frame.

on a very large data set is one preferable for speed?  one for memory use?

I tend to index using $ operators often and if I should quit let me know!!


Thanks,

Justin
> Logically i selected 2 rows with X[,"c"]==1. Than i want to
replace in that
> rows its own data from "d" to "b" with
X[,"b"]<-X[,"d"]. What is wrong?
> 
> --
> View this message in context:
http://r.789695.n4.nabble.com/conditional-data-replace-recode-change-or-whatsoever-tp3714715p3715218.html
> Sent from the R help mailing list archive at Nabble.com.
> 
>