在R中嵌套if else语句，每个单元格中包含多个字符串

Question

I would like to do an if else statement with multiple conditions. I have two data frames, first one looks like this:

prefix <- "sample"
suffix <- seq(1:100)
id <- paste(prefix, suffix, sep="")
indv_df <- data.frame(id, count = matrix(ncol=1, nrow=100))

And the first 15 rows of indv_df looks like this:

           id count
1     sample1    NA
2     sample2    NA
3     sample3    NA
4     sample4    NA
5     sample5    NA
6     sample6    NA
7     sample7    NA
8     sample8    NA
9     sample9    NA
10   sample10    NA
11   sample11    NA
12   sample12    NA
13   sample13    NA
14   sample14    NA
15   sample15    NA

The second table called row1 that I have looks like this:

 Hom <- paste("sample2", "sample3", "sample4", sep=",")
 Het <- paste("sample5", "sample6", "sample7", sep=",")
 Missing <- paste("sample10", "sample11", sep=",")
 row1 <- data.frame(Hom, Het, Missing)

looks like this:

                      Hom                     Het           Missing
1 sample2,sample3,sample4 sample5,sample6,sample7 sample10,sample11

I am trying to do if else statement that if the first row's id does not match any of the second table's content, write "0" in the first table's first row, second column. This is what I tried but didn't work, which I am not too surprised since this is my first if else statement. I know it should be straight forward but I tried a few different methods none worked

> if(grep(indv_df$id[1], row1$Hom)){
+   apply(indv_df[1,2]=="2")
+ } else if(grep(indv_df$id[1], row1$Het)){
+   apply(indv_df[1,2]=="1")
+ } else if(grep(indv_df$id[1], row1$Missing)){
+   apply(indv_df[1,2]=="missing")
+ } else (apply(indv_df[1,2]=="0"))

this is the error message I got:

Error in if (grep(indv_df$id[1], row1$Hom)) { : 
  argument is of length zero

The real dataset has 4 million rows in the second data.frame, so I am just testing the first step..... once I get through this I will try to do that in a loop for all rows. :D Thank you for all the help in advance.

Answer 1

A few issues that might be impacting you. Your eventual result will have a character column for the count column. It's best to do this up front, and would be clearer than how you currently have it done.

indv_df <- data.frame(id, count = NA_character_)

However, using your data.frame as you constructed it, I would approach this not by a series of if statements, but by subsetting. In addition, you have lines like apply(indv_df[1,2]=="missing") . This is wrong for a few reasons. indv_df[1,2] is more matrix syntax, and returns an NA. Then you check for identity with == , rather than assigning a value.

Here is a solution using data.frame syntax, and the stringr library.

library(stringr)
indv_df$count <- as.character(indv_df$count)
indv_df$count <- "0"
indv_df[indv_df$id %in% unlist(str_split(as.character(row1$Hom), ",")),]$count <- "2"
indv_df[indv_df$id %in% unlist(str_split(as.character(row1$Het), ",")),]$count <- "1"
indv_df[indv_df$id %in% unlist(str_split(as.character(row1$Missing), ",")),]$count <- "missing"
#          id   count
# 1   sample1       0
# 2   sample2       2
# 3   sample3       2
# 4   sample4       2
# 5   sample5       1
# 6   sample6       1
# 7   sample7       1
# 8   sample8       0
# 9   sample9       0
# 10 sample10 missing
# 11 sample11 missing
# 12 sample12       0
# 13 sample13       0
# 14 sample14       0
# 15 sample15       0

Personally, I prefer data.table syntax for this.

library(data.table)
library(stringr)
setDT(indv_df)
setDT(row1)
indv_df[, count := as.character(count)]
indv_df[, count := "0"]
indv_df[id %in% unlist(str_split(as.character(row1$Hom), ",")), count := "2"]
indv_df[id %in% unlist(str_split(as.character(row1$Het), ",")), count := "1"]
indv_df[id %in% unlist(str_split(as.character(row1$Missing), ",")), count := "missing"]
#           id   count
#  1:  sample1       0
#  2:  sample2       2
#  3:  sample3       2
#  4:  sample4       2
#  5:  sample5       1
#  6:  sample6       1
#  7:  sample7       1
#  8:  sample8       0
#  9:  sample9       0
# 10: sample10 missing
# 11: sample11 missing
# 12: sample12       0
# 13: sample13       0
# 14: sample14       0
# 15: sample15       0