Question

很抱歉，如果这是一个新手问题，但我在stackoverflow上找不到此特定问题的答案。我有一个（很大）固定宽度的数据文件，如下所示： simplefile.txt

[['This', 'is', 'the', '1st', 'line', '!'], ['This', 'is', 'the', '2nd', 'line', '.'], ['This', 'is', 'the', '3rd', 'line', '?'], ['This', 'is', 'the', '4th', 'line'], ['This', 'is', 'the', '5th', 'line', '.']]

我正在尝试使用如下所示的终端（该文件对于Excel而言太大）将其转换为csv文件：

ratno      fdate ratname                        typecode country        
12346 31/12/2010 HARTZ                              4    UNITED STATES
12444 31/12/2010 CHRISTIE                           5    UNITED STATES
12527 31/12/2010 HILL AIR                           4    UNITED STATES
15000 31/12/2010 TOKUGAVA  INC.                     5    JAPAN
37700 31/12/2010 HARTLAND                           1    UNITED KINGDOM
37700 31/12/2010 WILDER                             1    UNITED STATES  
18935 31/12/2010 FLOWERS FINAL SERVICES INC         5    UNITED STATES
37700 31/12/2010 MAPLE CORPORATION                  1    CANADA
48614 31/12/2010 SERIAL MGMT  L.P.                  5    UNITED STATES
 1373 31/12/2010 AMORE MGMT GROUP N A               1    UNITED STATES

我在此站点上进行了一些挖掘，发现了一种可能的解决方案，该解决方案依赖于ratno,fdate,ratname,typecode,country 12346,31/12/2010,HARTZ,4,UNITED STATES 12444,31/12/2010,CHRISTIE,5,UNITED STATES 12527,31/12/2010,HILL AIR,4,UNITED STATES 15000,31/12/2010,TOKUGAVA INC.,5,JAPAN 37700,31/12/2010,HARTLAND,1,UNITED KINGDOM 37700,31/12/2010,WILDER,1,UNITED STATES 18935,31/12/2010,FLOWERS FINAL SERVICES INC,5,UNITED STATES 37700,31/12/2010,MAPLE CORPORATION,1,CANADA 48614,31/12/2010,SERIAL MGMT L.P.,5,UNITED STATES 1373,31/12/2010,AMORE MGMT GROUP N A,1,UNITED STATES shell命令：

awk

但是，当我在终端中执行上述命令时，它会无意间在所有空白中插入逗号，而这些空白应该放在单个字段中。以上执行的结果如下：

awk -v FIELDWIDTHS="5 11 31 9 16" -v OFS=',' '{$1=$1;print}'   "simpletestfile.txt"

如何避免在限定的字段宽度之外的空白中插入逗号？谢谢！

Answer 1

您的尝试很好，但是需要FIELDWDTHS内置变量使用gawk（gnu awk）。使用gawk：

$ gawk -v FIELDWIDTHS="5 11 31 9 16" -v OFS=',' '{$1=$1;print}' file

ratno,      fdate, ratname                       , typecode, country
12346, 31/12/2010, HARTZ                         ,     4   , UNITED STATES
12444, 31/12/2010, CHRISTIE                      ,     5   , UNITED STATES
12527, 31/12/2010, HILL AIR                      ,     4   , UNITED STATES

假设您不需要多余的空格，可以改为：

$ gawk -v FIELDWIDTHS="5 11 31 9 16" -v OFS=',' '{for (i=1; i<=NF; ++i) gsub(/^ *| *$/, "", $i)}1' file
ratno,fdate,ratname,typecode,country
12346,31/12/2010,HARTZ,4,UNITED STATES
12444,31/12/2010,CHRISTIE,5,UNITED STATES
12527,31/12/2010,HILL AIR,4,UNITED STATES

如果您没有gnu awk，则可以使用以下方法获得相同的结果：

$ awk -v fieldwidths="5 11 31 9 16" '
BEGIN { OFS=","; split(fieldwidths, widths) }
{
    rec = $0
    $0 = ""
    start = 1;
    for (i=1; i<=length(widths); ++i) {
        $i = substr(rec, start, widths[i])
        gsub(/^ *| *$/, "", $i)
        start += widths[i]
    }
}1' file

ratno,fdate,ratname,typecode,country
12346,31/12/2010,HARTZ,4,UNITED STATES
12444,31/12/2010,CHRISTIE,5,UNITED STATES
12527,31/12/2010,HILL AIR,4,UNITED STATES

Answer 2

这里的perl很方便：

perl -nE '                                     # read this bottom to top
    say join ",", 
        map {s/^\s+|\s+$//g; $_}               # trim leading/trailing whitespace
        /^(.{5}) (.{10}) (.{30}) (.{8}) (.*)/  # extract the fields
' simplefile.txt

ratno,fdate,ratname,typecode,country
12346,31/12/2010,HARTZ,4,UNITED STATES
12444,31/12/2010,CHRISTIE,5,UNITED STATES
12527,31/12/2010,HILL AIR,4,UNITED STATES
15000,31/12/2010,TOKUGAVA  INC.,5,JAPAN
37700,31/12/2010,HARTLAND,1,UNITED KINGDOM
37700,31/12/2010,WILDER,1,UNITED STATES
18935,31/12/2010,FLOWERS FINAL SERVICES INC,5,UNITED STATES
37700,31/12/2010,MAPLE CORPORATION,1,CANADA
48614,31/12/2010,SERIAL MGMT  L.P.,5,UNITED STATES
1373,31/12/2010,AMORE MGMT GROUP N A,1,UNITED STATES

尽管如此，对于正确的CSV，我们需要对包含逗号或引号的字段保持谨慎。如果我对文件的内容不太安全，可以使用以下map块：

map {s/^\s+|\s+$//g; s/"/""/g; qq("$_")}

输出

"ratno","fdate","ratname","typecode","country"
"12346","31/12/2010","HARTZ","4","UNITED STATES"
"12444","31/12/2010","CHRISTIE","5","UNITED STATES"
"12527","31/12/2010","HILL AIR","4","UNITED STATES"
"15000","31/12/2010","TOKUGAVA  INC.","5","JAPAN"
"37700","31/12/2010","HARTLAND","1","UNITED KINGDOM"
"37700","31/12/2010","WILDER","1","UNITED STATES"
"18935","31/12/2010","FLOWERS FINAL SERVICES INC","5","UNITED STATES"
"37700","31/12/2010","MAPLE CORPORATION","1","CANADA"
"48614","31/12/2010","SERIAL MGMT  L.P.","5","UNITED STATES"
"1373","31/12/2010","AMORE MGMT GROUP N A","1","UNITED STATES"

无法将固定宽度的文件转换为csv

2 个答案: