Unix 按范围排序的数字序列

我有一个巨大的文本文件，每行都有排序的数字。我想使用sed或awk将其转换为范围块。 例如：

1
2
3
5
6
9
11
12

应转化为：

1-3
5-6
9
11-12

使用sed进行这种转换合理吗？

Perl解决方案：

perl -lne 'sub out { print $first == $prev ? $first : "$first-$prev" } # Produce the output.
           if ($_ != $prev+1 and defined $first) {                     # Current number not in seq.
               out();                                                  # Output the previous seq.
               $first = $_;                                            # Start a new one.
           }
           $prev  = $_;                                                # Remember the current number.
           $first = $_ if !defined $first;                             # Initialize for the first line.
           }{ out()                                                    # Output the last sequence.
          ' input.txt

awk 'NR==1 {printf $0} p+1!=$0 {printf "-"p"\n"$0} {p=$0} END {print "-"$0}' file
1-3
5-6
9-9
11-12

---

以下是一个

awk

解决方案：

perl -lne 'sub out { print $first == $prev ? $first : "$first-$prev" } # Produce the output.
           if ($_ != $prev+1 and defined $first) {                     # Current number not in seq.
               out();                                                  # Output the previous seq.
               $first = $_;                                            # Start a new one.
           }
           $prev  = $_;                                                # Remember the current number.
           $first = $_ if !defined $first;                             # Initialize for the first line.
           }{ out()                                                    # Output the last sequence.
          ' input.txt

awk 'NR==1 {printf $0} p+1!=$0 {printf "-"p"\n"$0} {p=$0} END {print "-"$0}' file
1-3
5-6
9-9
11-12

这假设所有行都有数字，并且每行的数字都会增加

---

使用

printf

awk 'NR==1 {printf "%s",$0} p+1!=$0 {printf "-%s\n%s",p,$0} {p=$0} END {print "-"$0}'

---

好了，伙计们，bash解决方案：

perl -lne 'sub out { print $first == $prev ? $first : "$first-$prev" } # Produce the output.
           if ($_ != $prev+1 and defined $first) {                     # Current number not in seq.
               out();                                                  # Output the previous seq.
               $first = $_;                                            # Start a new one.
           }
           $prev  = $_;                                                # Remember the current number.
           $first = $_ if !defined $first;                             # Initialize for the first line.
           }{ out()                                                    # Output the last sequence.
          ' input.txt

awk 'NR==1 {printf $0} p+1!=$0 {printf "-"p"\n"$0} {p=$0} END {print "-"$0}' file
1-3
5-6
9-9
11-12

A=`head-n1 data.txt`
B=一美元
读行时
做
如果[[`expr$LINE-$B`！=1]]
然后
如果[[“$A”！=“$B”]]
然后
回音“-$B”
其他的
回声
fi
A=美元行
回声-n“$A”
fi
B=美元行
done

sed-to-doculd会让人很痛苦，但因为sed无法计数，所以很难用awk来完成这项工作，而且CPU非常庞大。谢谢你们，awk也很好。我为awk添加了标签。谢谢@Jotne，太完美了！只是想分享一些细节。输入文件有22亿行，输出只有57k…并且在38分钟内终止。真棒：）