[R] get is.na & !is.na count of various combinations of columns

[arun]

Hi,
I just noticed that you also have all is.na() or !is.na() subsets.  To do that:
mydatatest <- structure(list(CaseID = structure(1:8, .Label = c("1605928", 
"1605943", "1605945", "1605947", "1605949", "1605951", "1605952", 
"1605953"), class = "factor"), Structure = structure(c(1L, 1L, 
3L, 2L, 2L, 3L, 3L, 4L), .Label = c("corp", "LLC", "prop", "LAC"
), class = "factor"), Poster = c(1, 1, 1, NA, NA, 1, 1, NA), 
    Records = c(1, 1, 1, 1, 1, NA, 1, NA), MWBW = c(NA, NA, 495.1, 
    NA, NA, NA, 425, NA), OT = c(NA, NA, NA, NA, 411.25, NA, 
    432.5, NA), CL = c(NA, NA, NA, 13.52, NA, NA, 14.72, NA)), .Names = c("CaseID", 
"Structure", "Poster", "Records", "MWBW", "OT", "CL"), row.names = c(NA, 
8L), class = "data.frame")



vec1 <- names(mydatatest)[-(1:2)]
res <- lapply(c(0,seq(length(vec1))),function(i) {x1 <- as.data.frame(combn(vec1,i),stringsAsFactors=FALSE); lapply(x1, function(x) {indx <- vec1 %in% x; nisna <- if(length(vec1[!indx]) > 0 ) paste(paste0("!is.na","(", vec1[!indx],")"),collapse=" & ");isna <- if(length(vec1[indx]) > 0 ) paste(paste0("is.na","(", vec1[indx],")"),collapse= " & "); nisna_isna <- gsub("^ & | & $","", paste(nisna, isna, sep= " & ")); subset(mydatatest, eval(parse(text=nisna_isna)))})})

names1 <-  unlist(lapply(c(0,seq(length(vec1))),function(i) {x1 <- as.data.frame(combn(vec1,i),stringsAsFactors=FALSE); unlist(lapply(x1, function(x) {indx <- vec1 %in% x; nisna <- if(length(vec1[!indx]) > 0 ) paste(paste0("!is.na","(", vec1[!indx],")"),collapse=" & ");isna <- if(length(vec1[indx]) > 0 ) paste(paste0("is.na","(", vec1[indx],")"),collapse= " & "); nisna_isna <- gsub("^ & | & $","", paste(nisna, isna, sep= " & "))}))}),use.names=FALSE)

res1 <- unlist(res,recursive=FALSE)
names(res1) <- names1
res2 <- res1[sapply(res1,nrow)!=0]


A.K.


On Wednesday, March 19, 2014 12:59 AM, arun <smartpink111 at yahoo.com> wrote:
If you want to identify the combination of columns:

names1 <-  unlist(lapply(seq(length(vec1)-1),function(i) {x1 <- as.data.frame(combn(vec1,i),stringsAsFactors=FALSE); unlist(lapply(x1, function(x) {indx <- vec1 %in% x; paste(paste(paste0("!is.na","(", vec1[!indx],")"),collapse=" & "), paste(paste0("is.na","(", vec1[indx],")"),collapse= " & "), sep= " & ") }))}),use.names=FALSE)


res1 <- unlist(res,recursive=FALSE)
 names(res1) <- names1

res1
 length(res1)
#[1] 30
res1[30]
#$`!is.na(Poster) & is.na(Records) & is.na(MWBW) & is.na(OT) & is.na(CL)`
#   CaseID Structure Poster Records MWBW OT CL
#6 1605951      prop      1      NA   NA NA NA
res2 <- res1[sapply(res1,nrow)!=0]
res2[4]
#$`!is.na(Poster) & !is.na(Records) & is.na(MWBW) & is.na(OT) & is.na(CL)`
#   CaseID Structure Poster Records MWBW OT CL
#1 1605928      corp      1       1   NA NA NA
#2 1605943      corp      1       1   NA NA NA

Hope this helps.


A.K.





On , arun <smartpink111 at yahoo.com> wrote:
Hi,
May be this helps:
res <-  lapply(seq(length(vec1)-1),function(i) {x1 <- as.data.frame(combn(vec1,i),stringsAsFactors=FALSE); lapply(x1, function(x) {indx <- vec1 %in% x; nisna <- paste(paste0("!is.na","(", vec1[!indx],")"),collapse=" & ");isna <- paste(paste0("is.na","(", vec1[indx],")"),collapse= " & ");subset(mydatatest, eval(parse(text=nisna)) & eval(parse(text=isna)))})})

A.K.





On Tuesday, March 18, 2014 11:45 PM, bcrombie <bcrombie at utk.edu> wrote:
I'm trying to count the number of combinations of columns containing data &
not containing data as described below.  I"m not sure how to do this in R
and need some help.

#get TRUE/FALSE count of various combinations of columns per CaseID or per
Structure
mydatatest <- data.frame (CaseID = c("1605928", "1605943", "1605945",
"1605947", "1605949", "1605951"),
                      Structure = c("corp", "corp", "prop", "LLC", "LLC",
"prop"),
                      Poster = c(1, 1, 1, NA, NA, 1),
                      Records = c(1, 1, 1, 1, 1, NA),
                      MWBW = c(NA, NA, 495.10, NA, NA, NA),
                      OT = c(NA, NA, NA, NA, 411.25, NA),
                      CL = c(NA, NA, NA, 13.52, NA, NA))

combo1 <- subset(mydatatest, !is.na(Poster & Records & MWBW & OT & CL))
combo2 <- subset(mydatatest, !is.na(Poster & Records & MWBW & OT) &
is.na(CL))
combo3 <- subset(mydatatest, !is.na(Poster & Records & MWBW) & is.na(OT &
CL))
combo4 <- subset(mydatatest, !is.na(Poster & Records) & is.na(MWBW & OT &
CL))
combo5 <- subset(mydatatest, !is.na(Poster) & is.na(Records & MWBW & OT &
CL))
combo6 <- subset(mydatatest, !is.na(Records) & is.na(Poster & MWBW & OT &
CL))
combo7 <- subset(mydatatest, !is.na(MWBW) & is.na(Poster & Records & OT &
CL))
combo8 <- subset(mydatatest, !is.na(OT) & is.na(Poster & Records & MWBW &
CL))
combo9 <- subset(mydatatest, !is.na(CL) & is.na(Poster & Records & MWBW &
OT))
etc.



--
View this message in context: http://r.789695.n4.nabble.com/get-is-na-is-na-count-of-various-combinations-of-columns-tp4687084.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.