linux - 将表格格式的输出转换为 bash 脚本中的逗号分隔表格

Question

我有类似的输出

No Type  Pid    Status  Cause Start Rstr  Err Sem Time Program          Cl  User         Action                    Table
-------------------------------------------------------------------------------------------------------------------------------
 0 DIA    10897 Wait          yes   no     0   0    0                                    NO_ACTION                           
 1 DIA    10903 Wait          yes   no     0   0    0                                    NO_ACTION                           
 2 DIA    10909 Wait          yes   no     0   0    0                                    NO_ACTION                           
 3 DIA    10916 Wait          yes   no     0   0    0                                    NO_ACTION                           
 4 DIA    10917 Wait          yes   no     0   0    0                                    NO_ACTION                           
 5 DIA     9061 Wait          yes   no     1   0    0                                    NO_ACTION                     

但我希望此表以逗号分隔，并且没有值的字段应该打印 null 而不是获取下一列的输出! 目前我收到以下输出。

NO=0,Type=DIA,Pid=10897,Status=Wait,Cause=yes,Start=no,Rstr=0,Err=0,Sem=0,Time=NO_ACTION,Program=,Cl=,User=,Action=,Table=
NO=1,Type=DIA,Pid=10903,Status=Wait,Cause=yes,Start=no,Rstr=0,Err=0,Sem=0,Time=NO_ACTION,Program=,Cl=,User=,Action=,Table=
NO=2,Type=DIA,Pid=10909,Status=Wait,Cause=yes,Start=no,Rstr=0,Err=0,Sem=0,Time=NO_ACTION,Program=,Cl=,User=,Action=,Table=
NO=3,Type=DIA,Pid=10916,Status=Wait,Cause=yes,Start=no,Rstr=0,Err=0,Sem=0,Time=NO_ACTION,Program=,Cl=,User=,Action=,Table=
NO=4,Type=DIA,Pid=10917,Status=Wait,Cause=yes,Start=no,Rstr=0,Err=0,Sem=0,Time=NO_ACTION,Program=,Cl=,User=,Action=,Table=
NO=5,Type=DIA,Pid=9061,Status=Wait,Cause=yes,Start=no,Rstr=1,Err=0,Sem=0,Time=NO_ACTION,Program=,Cl=,User=,Action=,Table=

我已经编写了一个脚本来执行相同的操作，但它不包括具有空值的列。

#!/bin/bash
sed 1,5d test.txt > temp.txt
input="temp.txt"
while IFS= read -r line
do
echo $line | awk 'BEGIN{FS=" ";OFS=","}{print "NO="$1,"Type="$2,"Pid="$3,"Status="$4,"Cause="$5,"Start="$6,"Rstr="$7,"Err="$8,"Sem="$9,"Time="$10,"Program="$11,"Cl="$12,"User="$13,"Action="$14,"Table="$15;}'
#echo "$line"
done < "$input"

Answer 1

我没有使用 awk 的经验，这显然可以使任务更快更短。<​​br/> 虽然这可以使用 bash 脚本完成，如下所示:

if [ "$#" -ne "2" ]
then
    echo "usage: <$0> input_file output_file"
    exit 1
fi

#input table file
input_file=$1
output_file=$2


#Get name for a temporary file by mktemp
temp_file=`mktemp headings_XXXXXX` 

#Store all headings separated by '\n' in a temporary file
sed -n '1p' $input_file | tr -s ' ' '\n' > $temp_file


headings=$(sed -n '1p' $input_file)

counter=0


#This loop would extract width of each column so that they can be given to cut as parameters
# like `cat filename | cut -b 3-8` would extract the entries in that column
while [ 1 ]
do
    upper_limit=${#headings}
    headings=${headings% [! ]*}
    lower_limit=${#headings}

    if [ "$upper_limit" = "$lower_limit" ]
    then
        limits_for_cut[$counter]=$(echo "1-${upper_limit}")
        counter=$( expr $counter + 1 )
        break
    fi

    lower_limit=$( expr $lower_limit + 1 )

    limits_for_cut[$counter]=$(echo "${lower_limit}-${upper_limit}")

    counter=$( expr $counter + 1 )

done


end_index=$( expr $counter - 1 )

no_of_lines=$( cat $input_file | wc -l )
no_of_lines=$( expr $no_of_lines - 2 ) #first 2 lines in file are for headings and dashes

on_line=$no_of_lines

#This loop will output all data to the specified file as comma separated
while [ $on_line -ne 0 ]
do
    counter=$end_index

    cat $temp_file |
        while read heading
        do
            tmp=$( expr $no_of_lines - $on_line + 1 + 2 )
            echo  "${heading}=`sed -n "${tmp}p" $input_file | cut -b ${limits_for_cut[$counter]} | sed 's/ //g'`," >> $output_file
            if [ $counter -eq 0 ]
            then
                break
            fi
            counter=$( expr $counter - 1 )
        done

    on_line=$( expr $on_line - 1 )
done    

echo `cat $output_file | tr -d '\n'` > $output_file

rm $temp_file

基本上，我们是用cut 命令来做的。

对于位于 3-8 之间的 header “type”，我们可以简单地这样做 cut -b 3-8 文件名。

我在 OSX 上运行了这个。您可能需要更改 cut 和 sed 语法 以适合您的机器。

如果此解决方案适合您，您应该尝试使用 awk 进行同样的操作，因为这会使它更快更短。<​​/p>