Difficulty importing text file with multiple different delimiters in R - r

Question

I'm having trouble figuring out how to import my data with multiple delimiters. The following is what my computer automatically saves into text file. The issue is that some of the results are printed with differently spaced delimiters. Some of the delimiters are colons (:) and others are multiple spaces with inconsistent length.
Each letter (B: to Z:) codes for some unique variable. For example:
B: Number of responses
C: Number of seconds, etc.
However, the information below "Z: 0.000", where the layout changes, is when the variables get subset. So,
A:
0: value1 value2 value3 value4
is referenced as:
A(0) = value1 (e.x. number of responses in the first trial)
A(1) = value2 (e.x. number of responses in the second trial)
A(2) = value3 (e.x. number of responses in the third trial)
A(3) = value4 (e.x. number of responses in the fourth trial)
Here, there are 4 "A" variables that each can carry unique values too.
Example of Text File:
Start Date: 05/20/18
End Date: 05/20/18
Subject: 1
Start Time: 16:23:11
End Time: 17:26:24
B: 7.000
C: 12000.000
D: 9500.000
E: 1.000
Q: 203.000
T: 1200.100
U: 218.000
W: 7.000
X: 347.000
Y: 0.000
Z: 0.000
A:
0: 1.000 0.000 0.000 0.000
F:
0: 11500.000 9500.000 13500.000 7500.000 15500.000
5: 5500.000 17500.000
I've tried a few methods, but they get stuck because the multiple delimiters issue. Let's assume "data" is the text file.
# This is the closest - some of the values are still not separated properly
temp <- read.delim2(file = "data", quote = ":", sep = "",)
# This one separate the information mostly correctly for the top half only
temp <- read.delim2(file = "data", sep = ":")
I eventually want a dataframe with labels in one column (StartDate, A(0), B, etc.) and values in the other (05/20/2018, 1, 7).

Answer 1

library(dplyr)
library(splitstackshape)
#read file
txt <- readLines("test.txt")
#Fix 'A:' rows
A_idx <- grep("A:", txt)
txt[A_idx] <- paste0(txt[A_idx], gsub("0:\\s+", "", txt[A_idx+1]))
txt <- txt[-(A_idx+1)]
#Fix 'F:' rows
F_idx <- grep("F:", txt)
txt[F_idx] <- paste0(txt[F_idx], paste(gsub("0:\\s+", "", txt[F_idx+1]),
gsub("5:\\s+", "", txt[F_idx+2])))
txt <- txt[-c(F_idx+1, F_idx+2)]
Now txt is in DCF format so it can be read using read.dcf
df <- data.frame(read.dcf(textConnection(txt)), stringsAsFactors = F) %>%
cSplit("A", " ") %>%
cSplit("F", " ")
Output is:
df
Start.Date End.Date Subject Start.Time End.Time B C D E Q T
1: 05/20/18 05/20/18 1 16:23:11 17:26:24 7.000 12000.000 9500.000 1.000 203.000 1200.100
U W X Y Z A_1 A_2 A_3 A_4 F_1 F_2 F_3 F_4 F_5 F_6 F_7
1: 218.000 7.000 347.000 0.000 0.000 1 0 0 0 11500 9500 13500 7500 15500 5500 17500
Sample data: test.txt contains
Start Date: 05/20/18
End Date: 05/20/18
Subject: 1
Start Time: 16:23:11
End Time: 17:26:24
B: 7.000
C: 12000.000
D: 9500.000
E: 1.000
Q: 203.000
T: 1200.100
U: 218.000
W: 7.000
X: 347.000
Y: 0.000
Z: 0.000
A:
0: 1.000 0.000 0.000 0.000
F:
0: 11500.000 9500.000 13500.000 7500.000 15500.000
5: 5500.000 17500.000
Start Date: 05/20/18
End Date: 05/20/18
... another block of data
Edit: If you want column A & F's index to start from 0
#read DCF data (i.e 'txt') using read.dcf
df <- data.frame(read.dcf(textConnection(txt)), stringsAsFactors = F)
#convert column A into wide format by splitting it into multiple columns
A_df <- data.frame(do.call(rbind, strsplit(as.character(df$A),'\\s+')), stringsAsFactors = F)
colnames(A_df) <- paste("A", sequence(ncol(A_df))-1, sep = "_")
#convert column F into wide format by splitting it into multiple columns
F_df <- data.frame(do.call(rbind, strsplit(as.character(df$F),'\\s+')), stringsAsFactors = F)
colnames(F_df) <- paste("F", sequence(ncol(F_df))-1, sep = "_")
#final data
final_df <- cbind(df[, !names(df) %in% c("A", "F")], A_df, F_df)
which gives
final_df
# Start.Date End.Date Subject Start.Time End.Time B C D E Q T U
#1 05/20/18 05/20/18 1 16:23:11 17:26:24 7.000 12000.000 9500.000 1.000 203.000 1200.100 218.000
# W X Y Z A_0 A_1 A_2 A_3 F_0 F_1 F_2 F_3 F_4
#1 7.000 347.000 0.000 0.000 1.000 0.000 0.000 0.000 11500.000 9500.000 13500.000 7500.000 15500.000
# F_5 F_6
#1 5500.000 17500.000

Answer 2

The good news is that your file does NOT have different delimiters. It is "Debian Control File" format. The whitespace marks continuous lines. See ?read.dcf Unfortunately, I cannot figure out if there is a way to parse .dcf including the semantics of continuous lines. But what the heck, once the data is in R, you can just clean it with library(tidyr)
x <- read.dcf("yoursourcefilename.txt")
y <– as.data.frame(x) # read.dcf reads in as matrix
z <- y %>%
separate("A", into = c("drop", "A0"), sep = "0:") %>%
separate("A0", into = c("drop", paste0("A0_val_", 1:4)), sep = "\\s{2,}") %>%
separate("F", into = c("drop", "F0"), sep = "0:") %>%
separate("F0", into = c("F0", "F5"), sep = "5:") %>%
separate("F0", into = c("drop", paste0("F0_val_", 1:5)), sep = "\\s{2,}") %>%
separate("F5", into = c("drop", paste0("F5_val_", 1:2)), sep = "\\s{2,}") %>%
select(-drop) %>% t() %>% as.data.frame()
z$V1 <- trimws(z$V1) # clean whatever whitespace is left
This will yield you a long dataframe:
dim(z)
[1] 27 1
Like so:
> z
V1
Start Date 05/20/18
End Date 05/20/18
Subject 1
Start Time 16:23:11
End Time 17:26:24
B 7.000
C 12000.000
D 9500.000
E 1.000
Q 203.000
T 1200.100
U 218.000
W 7.000
X 347.000
Y 0.000
Z 0.000
F5_val_1 5500.000
F5_val_2 17500.000
F0_val_1 11500.000
F0_val_2 9500.000
F0_val_3 13500.000
F0_val_4 7500.000
F0_val_5 15500.000
A0_val_1 1.000
A0_val_2 0.000
A0_val_3 0.000
A0_val_4 0.000
I am not sure this is the most efficient to work with the data (not a tidy format), but sounds like this is what you wanted?