How to properly import a semicolon

I have been trying to import the following semi-colon separated file:

# word  len;freq;mean;sens;npos;u;orthon;freqn;bgp  WN000000
fiber   "5;8.671;1;5;1;0;5;6.1;0;-1"    10000000
clad    "4;6.78;2;2;1;1;8;7.84;2026;-1" 10000000
tucker  "6;8.103;2;3;2;0.91829583405449;7;5.5;4547;-1"  10000000

I tried both read.csv and data.table::fread , but with no luck. read.csv recognizes some of headers and the actual values are all under the first column:

 X..word.len freq mean sens npos  u orthon freqn bgp.WN000000
1 fibert5;8.671;1;5;1;0;5;6.1;0;-1t10000000   NA   NA   NA   NA NA     NA    NA           NA
2 cladt4;6.78;2;2;1;1;8;7.84;2026;-1t10000000   NA   NA   NA   NA NA     NA    NA           NA
3 tuckert6;8.103;2;3;2;0.91829583405449;7;5.5;4547;-1t10000000   NA   NA   NA   NA NA     NA    NA           NA

fread recognizes the first column, but merges all the others into one.

  X..word    len.freq.mean.sens.npos.u.orthon.freqn.bgp WN000000
1   fiber                    5;8.671;1;5;1;0;5;6.1;0;-1 10000000
2    clad                 4;6.78;2;2;1;1;8;7.84;2026;-1 10000000
3  tucker  6;8.103;2;3;2;0.91829583405449;7;5.5;4547;-1 10000000

Can anyone help out?


Here is what I try (and edited based on David Arenburg 's comment).

Read and process the header line first; then read the remaining lines while skipping the first line:

library(data.table)

header <- strsplit(readLines('test.txt', n = 1), 's+')[[1]][-1]
res <- fread('test.txt', skip = 1, header = FALSE)
setnames(res, 1:3, header)

res[, strsplit(header[2], ';')[[1]] :=
        tstrsplit(get(header[2]), ';', type.convert = TRUE, fixed = TRUE)[-10]]

res[, header[2] := NULL]

#      word WN000000 len  freq mean sens npos         u orthon freqn  bgp
# 1:  fiber 10000000   5 8.671    1    5    1 0.0000000      5  6.10    0
# 2:   clad 10000000   4 6.780    2    2    1 1.0000000      8  7.84 2026
# 3: tucker 10000000   6 8.103    2    3    2 0.9182958      7  5.50 4547

It should be noted that there are 9 items seperated by ; in second column of the input file, but the following rows have 10 ; -seperated items.


As far as I can tell, you've basically got two separators working there. It might be worth handling this manually, rather than using a standard reading function.

Read the file in, get rid of the quotations, then separate by both space and semicolon delimiters.

library( magrittr )
input <- readLines( "~/Desktop/Untitled" ) %>%
    gsub( '"|^# +', "", . ) %>%
    strsplit( " +|;" )

Convert everything except the first row to a data frame.

input[-1] %>%
    do.call( what = rbind ) %>%
    as.data.frame()

Then use the first row as column names. We need to add an extra here, since your header doesn't have enough names to cover your data.

names( df ) <- c( input[[1]][1:10], "fill.col", input[[1]][11] )

Result:

> df
    word len  freq mean sens npos                u orthon freqn  bgp fill.col WN000000
1  fiber   5 8.671    1    5    1                0      5   6.1    0       -1 10000000
2   clad   4  6.78    2    2    1                1      8  7.84 2026       -1 10000000
3 tucker   6 8.103    2    3    2 0.91829583405449      7   5.5 4547       -1 10000000
链接地址: http://www.djcxy.com/p/96270.html

上一篇: JSlint错误'不要在循环中创建函数。' leads to question about Javascript

下一篇: 如何正确导入分号