按指定的列长度将ASCII数据读入R

Question

I have a large ASCII file that looks something like this if you open it in a text editor:

11112223423 4434 555534 5533 54534 5354 5532 434 4 43434 23424234 34 4534 34453 345345345 345344 344 43423453453 43444 99098 234090 4354550 345399 43453 9900 4

I have been given a mapping of the columns. For example: The first variable sits in columns 1-9. The second column sits in 104-105. And so on.

Is there an easy way to read this type of data into R so that I end up with a data.frame?

Thanks for the help!

Answer 1

I've used the standard read.fwf() for this kind of thing.

I also like read_fwf() from the readr package. For example:

#create some dummy fixed-width-field data
fixed_width_data <- "line1  field1 datafield2 dataetc\nline2  field1 datafield2 dataetc\n"

#specify the data columns
field_info <- fwf_widths(c(7, 11, 11, 3), col_names=c("line_number", "field1", "field2", "fieldn"))

#read it in
parsed <- read_fwf(fixed_width_data,  field_info)

To specify start/end positions for the columns of data, you can use fwf_positions() instead of fwf_widths() :

#create some dummy fixed-width-field data
fixed_width_data2 <- "line1  field1 datafield2 dataTEXT TO SKIPetc\nline2  field1 datafield2 dataTEXT TO SKIPetc\n"

#specify the data columns using start and end positions
field_info2 <- fwf_positions(start=c(1, 8, 19, 42), end=c(5, 18, 29, 44), col_names=c("line_number", "field1", "field2", "fieldn"))

#read it in
parsed2 <- read_fwf(fixed_width_data2,  field_info2)

Answer 2

You can do this in base R using read.fwf (fixed width fields) I wrote a file with your single line of input and got:

FullFile = read.fwf("Test.txt", widths=c(9,94,2))
Interesting = FullFile[,c(1,3)]
Interesting
         V1 V3
1 111122234 42

Note that I am reading the columns to skip into a variable and then just discarding that variable.