CODEWITHSUNDEEP
rdplyrapply
maximusdooku
maximusdooku

Reputation: 5522

What is a dplyr equivalent for using which function?

I have a dataframe 1: 

df1<- structure(list(`gaugelist[1:20, ]` = c(1094000L, 1094000L, 1094000L, 
1100600L, 1100600L, 1100600L, 1100600L, 1100600L, 1100600L, 1100600L, 
1100600L, 1100600L, 1100600L, 1100600L, 1096000L, 1100600L, 1100600L, 
1100600L, 1100600L, 1100600L)), .Names = "gaugelist[1:20, ]", row.names = c(NA, 
-20L), class = "data.frame")

And a second list:

df2 <- c("ts.01094000.crest.csv", "ts.01100600.crest.csv")

I want to make another column in df1 based on matching row values:

df1$test <- apply(df1, 1, function(x) df2[which(grepl(x,df2))])
  1. What would be the dplyr way of making this work? I couldn't make rowwise work.
  2. How can I make this work if df1 has large number of columns and I want to just one new column 'test' based on same rule.

Upvotes: 0

Views: 1303

Answers (1)

Weihuang Wong
Weihuang Wong

Reputation: 13118

This worked for me:

df1 <- setNames(df1, "x") # rename column to `x` for brevity
df1 %>% rowwise() %>%
 mutate(test = df2[grep(x,df2)[1]])
# Source: local data frame [20 x 2]
# Groups: <by row>

# # A tibble: 20 × 2
#          x                  test
#      <int>                 <chr>
# 1  1094000 ts.01094000.crest.csv
# 2  1094000 ts.01094000.crest.csv
# 3  1094000 ts.01094000.crest.csv

Explanation. You did not explain why you "couldn't make rowwise work", but based on my own experience, you may have gotten this message

Error: incompatible size (0), expecting 1 (the group size) or 1

if you use an expression like df2[which(grepl(x,df2))] in mutate, which occurs because one (or more) of the returned values has length 0 (i.e. no match was found in df2). Using mutate(test = df2[grep(x,df2)[1]]) addresses this problem. In particular, note that df2[grep(x,df2)[1]] returns one and only one value.

Upvotes: 2

Related Questions