Changing multiple colnames simultaneously replacing data with NA - r

I'm trying to rename multiple column names of a data frame in which the columns contain more than a single type, the columns are a factor class.
col1 col2 col3 col4 col5 col6
a b c a b a
1 5 8 2 2 5
conditional on an entry in a row:
colnames(df)[which(df[1,]=="b " )]<-"new_colname"
Ideally producing something like:
col1 new_colname col3 col4 new_colname.2 col6
a b c a b a
1 5 8 2 2 5
But when I do this all the columns that are renamed have their data replaced with NAs, producing:
col1 col2 col3
NA NA NA
NA NA NA
Does anyone know why this would happen?
The

Suppose, the dataset columns are all "factor" class, convert the columns to "character" class.
df[] <- lapply(df, as.character)
In case, there are leading/lagging spaces, use str_trim to remove those spaces,
library(stringr)
df[] <- lapply(df, str_trim)
Change the column names based on the conditions mentioned, and use make.names for creating unique names for those duplicated column names.
names(df)[df[1,]=='b'] <- 'new_colname'
names(df) <- make.names(names(df), unique=TRUE)
df
# col1 new_colname col3 col4 new_colname.1 col6
#1 a b c a b a
#2 1 5 8 2 2 5
data
df <- structure(list(col1 = structure(c(2L, 1L), .Label = c("1", "a"
), class = "factor"), col2 = structure(c(2L, 1L), .Label = c("5",
"b"), class = "factor"), col3 = structure(c(2L, 1L), .Label = c("8",
"c"), class = "factor"), col4 = structure(c(2L, 1L), .Label = c("2",
"a"), class = "factor"), col5 = structure(c(2L, 1L), .Label = c("2",
"b"), class = "factor"), col6 = structure(c(2L, 1L), .Label = c("5",
"a"), class = "factor")), .Names = c("col1", "col2", "col3",
"col4", "col5", "col6"), row.names = c(NA, -2L), class = "data.frame")

In the end fixed this by naming them using a different method, using a for loop:
for(i in 1:length(df)){colnames(df)[i]<-paste("df", df[1,i],df[3,i], eval(i) ,sep="_" )}
This would probably not be feasible for an extremely large dataset, so if anyone knows how one might do this another way please post an answer.

Related

How to collapse rows by identical values in a column

Good evening,
I have a two columns tab separated .txt file, as the following:
number letter
1 a
1 b
2 a
2 b
3 b
I would like to collapse rows where the column "number" has identical value, by creating a comma separated value in the corresponding column "letter".
In other words, this should be the output:
number letter
1 a,b
2 a,b
3 b
I have looked up the web but I did not find an actual solution.
Thank you in advance,
Giuseppe
We can use aggregate in base R
aggregate(letter ~ number, df1, FUN = paste, collapse=",")
-output
# number letter
#1 1 a,b
#2 2 a,b
#3 3 b
Or with tidyverse
library(dplyr)
library(stringr)
df1 %>%
group_by(number) %>%
summarise(letter = str_c(letter, collapse=","))
data
df1 <- structure(list(number = c(1L, 1L, 2L, 2L, 3L), letter = c("a",
"b", "a", "b", "b")), class = "data.frame", row.names = c(NA,
-5L))
We can also combine aggregate() with toString:
#Code
newdf <- aggregate(letter~.,df,toString)
Output:
number letter
1 1 a, b
2 2 a, b
3 3 b
Some data:
#Data
df <- structure(list(number = c(1L, 1L, 2L, 2L, 3L), letter = c("a",
"b", "a", "b", "b")), class = "data.frame", row.names = c(NA,
-5L))

R: Combining dataset and lookup-table to extract value to new colume

I want to combine to dataframes, df1 with 15.000 obs and df2 consisting of 2.3 mill. I'm trying to match values, if df1$col1 == df2$c1, AND df1$col2 == df2$c2, then insert value from df2$dummy, to df1$col3. If no match in both, do nothing. All are 8 digits, except df2$dummy, which is a dummy of 0 or 1.
df1 col1 col2 col3
1 25382701 65352617 -
2 22363658 45363783 -
3 20019696 23274747 -
df2 c1 c2 dummy
1 17472802 65548585 1
2 20383829 24747473 0
3 20019696 23274747 0
4 01382947 21930283 1
5 22123425 65382920 0
In the example the only match is row 3, and the value 0 from the dummy column should be inserted in col3 row3.
I've tried to make a look-up table, a function of for and if, but not found a solution when requiring matches in two dataframes. (No need to say I guess, but I'm new to R and programming..)
We can use a join in data.table
library(data.table)
df1$col3 <- NULL
setDT(df1)[df2, col3 := i.dummy, on = .(col1 = c1, col2 = c2)]
df1
# col1 col2 col3
#1: 25382701 65352617 NA
#2: 22363658 45363783 NA
#3: 20019696 23274747 0
data
df1 <- structure(list(col1 = c(25382701L, 22363658L, 20019696L), col2 = c(65352617L,
45363783L, 23274747L), col3 = c("-", "-", "-")), class = "data.frame", row.names = c("1",
"2", "3"))
df2 <- structure(list(c1 = c(17472802L, 20383829L, 20019696L, 1382947L,
22123425L), c2 = c(65548585L, 24747473L, 23274747L, 21930283L,
65382920L), dummy = c(1L, 0L, 0L, 1L, 0L)), class = "data.frame",
row.names = c("1",
"2", "3", "4", "5"))

Replace a subset of data frame

I have a data frame with some error
T item V1 V2
1 a 2 .1
2 a 5 .8
1 b 1 .7
2 b 2 .2
I have another data frame with corrections for items concerning V1 only
T item V1
1 a 2
2 a 6
How do I get the final data frame? Should I use merge or rbind. Note: actual data frames are big.
An option would be a data.table join on the 'T', 'item' and assigning the 'V1' with the the corresponding 'V1' column (i.V1) from the second dataset
library(data.table)
setDT(df1)[df2, V1 := i.V1, on = .(T, item)]
df1
# T item V1 V2
#1: 1 a 2 0.1
#2: 2 a 6 0.8
#3: 1 b 1 0.7
#4: 2 b 2 0.2
data
df1 <- structure(list(T = c(1L, 2L, 1L, 2L), item = c("a", "a", "b",
"b"), V1 = c(2L, 5L, 1L, 2L), V2 = c(0.1, 0.8, 0.7, 0.2)),
class = "data.frame", row.names = c(NA, -4L))
df2 <- structure(list(T = 1:2, item = c("a", "a"), V1 = c(2L, 6L)),
class = "data.frame", row.names = c(NA,
-2L))
This should work -
library(dplyr)
df1 %>%
left_join(df2, by = c("T", "item")) %>%
mutate(
V1 = coalesce(as.numeric(V1.y), as.numeric(V1.x))
) %>%
select(-V1.x, -V1.y)

Reduce() in R over similar variable names causing error

I have 19 nested lists generated from a lapply and split operation.
These lists are in the form:
#list1
Var col1 col2 col3
A 2 3 4
B 3 4 5
#list2
Var col1 col2 col3
A 5 6 7
B 5 4 4
......
#list19
Var col1 col2 col3
A 3 6 7
B 7 4 4
I have been able to merge the lists with
merge.all <- function(x, y) merge(x, y, all=TRUE, by="Var")
out <- Reduce(merge.all, DataList)
I am however getting an error due to the similarity in the names of the other columns.
How can I concatenate the name of the list to the variable names so that I get something like this:
Var list1.col1 list1.col2 list1.col3 .......... list19.col3
A 2 3 4 7
B 3 4 5 .......... 4
I'm really sure somebody will come up with a much, much better solution. However, if you're after a quick and dirty solution, this seems to work.
My plan was to simply change the column names prior to merging.
#Sample Data
df1 <- data.frame(Var = c("A","B"), col1 = c(2,3), col2 = c(3,4), col3 = c(4,5))
df2 <- data.frame(Var = c("A","B"), col1 = c(5,5), col2 = c(6,4), col3 = c(7,5))
df19 <- data.frame(Var = c("A","B"), col1 = c(3,7), col2 = c(6,4), col3 = c(7,4))
mylist <- list(df1, df2, df19)
names(mylist) <- c("df1", "df2", "df19") #just manually naming, presumably your list has names
## Change column names by pasting name of dataframe in list with standard column names. - using ugly mix of `lapply` and a `for` loop:
mycolnames <- colnames(df1)
mycolnames1 <- lapply(names(mylist), function(x) paste0(x, mycolnames))
for(i in 1:length(mylist)){
colnames(mylist[[i]]) <- mycolnames1[[i]]
colnames(mylist[[i]])[1] <- "Var" #put Var back in so you can merge
}
## Merge
merge.all <- function(x, y)
merge(x, y, all=TRUE, by="Var")
out <- Reduce(merge.all, mylist)
out
# Var df1col1 df1col2 df1col3 df2col1 df2col2 df2col3 df19col1 df19col2 df19col3
#1 A 2 3 4 5 6 7 3 6 7
#2 B 3 4 5 5 4 5 7 4 4
There you go - it works but is very ugly.
To set the data frame names unique, you could use a function to set all list names that are not the merging variable to unique names.
resetNames <- function(x, byvar = "Var") {
asrl <- as.relistable(lapply(x, names))
allnm <- names(unlist(x, recursive = FALSE))
rpl <- replace(allnm, unlist(asrl) %in% byvar, byvar)
Map(setNames, x, relist(rpl, asrl))
}
Reduce(merge.all, resetNames(dlist))
# Var list1.col1 list1.col2 list1.col3 list2.col1 list2.col2 list2.col4 list3.col1
#1 A 2 3 4 5 6 7 3
#2 B 3 4 5 5 4 4 7
# list3.col2 list3.col3 list4.col1 list4.col2 list4.col3
#1 6 7 3 6 7
#2 4 4 4 5 6
when run your list with an added data frame there are no warnings. And there's always data table. Its merge method does not return a warning for duplicated column names.
library(data.table)
Reduce(merge.all, lapply(dlist, as.data.table))
Another option is to check the names as the data enters the function, change them there, and then you can avoid the warning. This isn't perfect but it works ok here.
merge.all <- function(x, y) {
m <- match(names(y)[-1], gsub("[.](x|y)$", "", names(x)[-1]), 0L)
names(y)[-1][m] <- paste0(names(y)[-1][m], "DUPE")
merge(x, y, all=TRUE, by="Var")
}
rm <- Reduce(merge.all, dlist)
names(rm)
# [1] "Var" "col1" "col2" "col3" "col1DUPE.x"
# [6] "col2DUPE.x" "col4" "col1DUPE.y" "col2DUPE.y" "col3DUPE.x"
# [11] "col1DUPE" "col2DUPE" "col3DUPE.y"
where dlist is
structure(list(list1 = structure(list(Var = structure(1:2, .Label = c("A",
"B"), class = "factor"), col1 = 2:3, col2 = 3:4, col3 = 4:5), .Names = c("Var",
"col1", "col2", "col3"), class = "data.frame", row.names = c(NA,
-2L)), list2 = structure(list(Var = structure(1:2, .Label = c("A",
"B"), class = "factor"), col1 = c(5L, 5L), col2 = c(6L, 4L),
col4 = c(7L, 4L)), .Names = c("Var", "col1", "col2", "col4"
), class = "data.frame", row.names = c(NA, -2L)), list3 = structure(list(
Var = structure(1:2, .Label = c("A", "B"), class = "factor"),
col1 = c(3L, 7L), col2 = c(6L, 4L), col3 = c(7L, 4L)), .Names = c("Var",
"col1", "col2", "col3"), class = "data.frame", row.names = c(NA,
-2L)), list4 = structure(list(Var = structure(1:2, .Label = c("A",
"B"), class = "factor"), col1 = 3:4, col2 = c(6L, 5L), col3 = c(7L,
6L)), .Names = c("Var", "col1", "col2", "col3"), row.names = c(NA,
-2L), class = "data.frame")), .Names = c("list1", "list2", "list3",
"list4"))

R paste0 2 columns if not NA

I would like to paste0 two columns if the element in one column is not NA.If one element of one columns is NA then keep the element of the other column only.
structure(list(col1 = structure(1:3, .Label = c("A", "B", "C"),
class = "factor"), col2 = c(1, NA, 3)), .Names = c("col1", "col2"),
class = "data.frame",row.names = c(NA, -3L))
# col1 col2
# 1 A 1
# 2 B NA
# 3 C 3
structure(list(col1 = structure(1:3, .Label = c("A", "B", "C"),
class = "factor"),col2 = c(1, NA, 3), col3 = c("A|1", "B", "C|3")),
.Names = c("col1", "col2", "col3"), row.names = c(NA,-3L),
class = "data.frame")
# col1 col2 col3
#1 A 1 A|1
#2 B NA B
#3 C 3 C|3
you can also do it with regular expressions:
df$col3 <- sub("NA\\||\\|NA", "", with(df, paste0(col1, "|", col2)))
That is, paste them in regular way and then replace any "NA|" or "|NA" with "". Note that | needs to be "double escaped" because it means "OR" in regexps, that's why the strange pattern NA\\||\\|NA means actually "NA|" OR "|NA".
As #Roland says, this is easy using ifelse (just translate the mental logic into a series of nested ifelse statements):
x <- transform(x,col3=ifelse(is.na(col1),as.character(col2),
ifelse(is.na(col2),as.character(col1),
paste0(col1,"|",col2))))
update: need as.character in some cases.
Try:
> df$col1 = as.character(df$col1)
> df$col3 = with(df, ifelse(is.na(col1),col2, ifelse(is.na(col2), col1, paste0(col1,'|',col2))))
> df
col1 col2 col3
1 A 1 A|1
2 B NA B
3 C 3 C|3
You could also do:
library(stringr)
df$col3 <- apply(df, 1, function(x)
paste(str_trim(x[!is.na(x)]), collapse="|"))
df
# col1 col2 col3
#1 A 1 A|1
#2 B NA B
#3 C 3 C|3

Resources