Extract data.frame column names from rows in data.frame - r

Question

I have a couple of data.frames which have approximately the same structure. For a reproducible example I created two sample dataframes df1 and df2.
df1 <- structure(list(sample = c(2L, 6L), data1 = c(56L, 78L), data2 = c(59L,
27L), data6 = c(90L, 28L), data1namet = structure(c(1L, 1L), .Label = "Sam1", class = "factor"),
data2namab = structure(c(1L, 1L), .Label = "Test2", class = "factor"),
dataame = structure(c(1L, 1L), .Label = "Ex3", class = "factor")), .Names = c("sample",
"data1", "data2", "data3", "data1namet", "data2namab", "dataame"
), class = "data.frame", row.names = c(NA, -2L))
df1
sample data1 data2 data3 data1namet data2namab dataame
1 2 56 59 90 Sam1 Test2 Ex3
2 6 78 27 28 Sam1 Test2 Ex3
df2 <- structure(list(sample = c(12L, 13L, 17L), data1 = c(56L, 78L,
3L), data2 = c(59L, 27L, 2L), datest = structure(c(1L, 1L,
1L), .Label = "Exa9", class = "factor"), dattestr = structure(c(1L,
1L, 1L), .Label = "cz1", class = "factor")), .Names = c("sample",
"data1", "data2", "datest", "dattestr"), class = "data.frame", row.names = c(NA,
-3L))
df2
sample data1 data2 datest dattestr
1 12 56 59 Exa9 cz1
2 13 78 27 Exa9 cz1
3 17 3 2 Exa9 cz1
The name of the data is saved in the columns after the data columns and I was wondering if there is a way I could restructure the data.frames (about 40 data.frames) that they contain the name of the data in their column name?
df1
sample Sam1 Test2 Ex3
1 2 56 59 90
2 6 78 27 28
and
df2
sample Exa9 cz1
1 12 56 59
2 13 78 27
3 17 3 2
EDIT
As I just realised I also have other columns after the data columns so that my input data looks like this
df1 <- structure(list(sample = c(2L, 6L), data1 = c(56L, 78L), data2 = c(59L,
27L), data3 = c(90L, 28L), data1namet = structure(c(1L, 1L), .Label = "Sam1", class = "factor"),
data2namab = structure(c(1L, 1L), .Label = "Test2", class = "factor"),
dataame = structure(c(1L, 1L), .Label = "Ex3", class = "factor"),
ma = c("Jay", "Jay")), .Names = c("sample", "data1", "data2",
"data3", "data1namet", "data2namab", "dataame", "ma"), row.names = c(NA,
-2L), class = "data.frame")
df1
sample data1 data2 data3 data1namet data2namab dataame ma
1 2 56 59 90 Sam1 Test2 Ex3 Jay
2 6 78 27 28 Sam1 Test2 Ex3 Jay
df2 <- structure(list(sample = c(12L, 13L, 17L), data1 = c(56L, 78L,
3L), data2 = c(59L, 27L, 2L), datest = structure(c(1L, 1L, 1L
), .Label = "Exa9", class = "factor"), dattestr = structure(c(1L,
1L, 1L), .Label = "cz1", class = "factor"), add = c(2, 2, 2)), .Names = c("sample",
"data1", "data2", "datest", "dattestr", "add"), row.names = c(NA,
-3L), class = "data.frame")
df2
sample data1 data2 datest dattestr add
1 12 56 59 Exa9 cz1 2
2 13 78 27 Exa9 cz1 2
3 17 3 2 Exa9 cz1 2
In this case the ma and add column are not part of the data and should be added at the end like this:
df1
sample Sam1 Test2 Ex3 ma
1 2 56 59 90 Jay
2 6 78 27 28 Jay
and
df2
sample Exa9 cz1 add
1 12 56 59 2
2 13 78 27 2
3 17 3 2 2

Answer 1

One could start by identifying which columns should be kept:
keep_col <- which(sapply(df2, is.numeric))
After that, some work is required to extract the new column names and to rename the corresponding columns in the data frame:
names <- df2[1,keep_col[-1] + length(keep_col)-1]
colnames(df2)[keep_col[-1]] <- as.character(unlist(names))
Finally, the dataframe can be reassembled by keeping only the desired columns:
df2 <- df2[,keep_col]
#> df2
# sample Exa9 cz1
#1 12 56 59
#2 13 78 27
#3 17 3 2
In order to use this transformation for several different dataframes, the code can be wrapped into a function:
summarize_table <- function(x){
keep_col <- which(sapply(x, is.numeric))
names <- x[1,keep_col[-1] + length(keep_col)-1]
colnames(x)[keep_col[-1]] <- as.character(unlist(names))
x <- x[,keep_col]
}
If the various dataframes are stored in a list, the function summarize_table() can be used with lapply() to obtain the results for each dataframe:
my_dfs <- list(df1,df2)
out <- lapply(my_dfs,summarize_table)
#> out
#[[1]]
# sample Sam1 Test2 Ex3
#1 2 56 59 90
#2 6 78 27 28
#
#[[2]]
# sample Exa9 cz1
#1 12 56 59
#2 13 78 27
#3 17 3 2
EDIT / ADDENDUM
The modified version below should be able to handle also the cases mentioned in the revised post:
summarize_tab2 <- function(x){
keep_col <- which(sapply(x, is.numeric))
first_block <- c(keep_col[1],keep_col[which(diff(keep_col)==1)])
add_col <- FALSE
if (2 * (length(keep_col) - 1) + 1 < ncol(x)) add_col <- TRUE
keep_col1 <- keep_col[1:length(first_block)]
names <- x[1,keep_col1[-1] + length(keep_col1) - 1]
colnames(x)[keep_col1[-1]] <- as.character(unlist(names))
df_t <- x[,keep_col]
if (add_col) df_t <- cbind(df_t, x[(2 * (ncol(df_t) - 1) + 2):ncol(x)])
return(df_t)
}
my_dfs <- list(df1, df2, df3, df4)
out <- lapply(my_dfs, summarize_tab2)
#> out
#[[1]]
# sample Sam1 Test2 Ex3 ma
#1 2 56 59 90 Jay
#2 6 78 27 28 Jay
#
#[[2]]
# sample Exa9 cz1 add
#1 12 56 59 2
#2 13 78 27 2
#3 17 3 2 2
#
#[[3]]
# sample Sam1 Test2 Ex3
#1 2 56 59 90
#2 6 78 27 28
#
#[[4]]
# sample Exa9 cz1
#1 12 56 59
#2 13 78 27
#3 17 3 2
Here the dataframes df3 and df4 are, respectively, the data frames df1and df2 of the original post.

Answer 2

The following should work:
library(plyr)
cols.to.rename <- grep('^data(.)$', colnames(df1))
cols.of.names <- max(cols.to.rename)+seq(1,length(cols.to.rename))
the.names <- lapply(df1[1,cols.of.names], as.character)
df1.mod <- df1
colnames(df1.mod)[cols.to.rename] <- the.names
df1.mod <- df1.mod[-cols.of.names]
It renames all dataX columns to the (first) value in the columns following the last dataX column. It then drops all name columns from the data frame.