unix - Unix 性能改进 - 可能正在使用 AWK

Question

我有两个文件 File1.txt（它有 6 列由管道分隔）和 File2.txt（它有 2 列由管道分隔）

文件1.txt

NEW|abcd|1234|10000000|Hello|New_value|
NEW|abcd|1234|20000000|Hello|New_value|
NEW|xyzq|5678|30000000|myname|New_Value|

文件2.txt

10000000|10000001>10000002>10000003>10000004
19000000|10000000>10000001>10000002>10000003>10000004
17000000|10000099>10000000>10000001>10000002>10000003>10000004
20000000|10000001>10000002>10000003>10000004>30000000
29000000|20000000>10000001>10000002>10000003>10000004

目标是针对 File1.txt 中的每一行，我必须选择第 4 列并且必须在 File2.txt 中搜索该值。如果在 File2.txt 中找到任何匹配项，那么我必须从 File2.txt 中提取所有行，但只提取第一列。

这可能会在目标文件中产生更多的记录。输出应如下所示（最后一列 123 来自固定变量）

NEW|abcd|1234|10000000|Hello|New_value|123    (this row comes as it matches 1st row & 4th column of File1.txt with 1st row of File2.txt)
NEW|abcd|1234|19000000|Hello|New_value|123    (this row comes as it matches 1st row & 4th column of File1.txt with 2nd row of File2.txt)
NEW|abcd|1234|17000000|Hello|New_value|123    (this row comes as it matches 1st row & 4th column of File1.txt with 3rd row of File2.txt)
NEW|abcd|1234|20000000|Hello|New_value|123    (this row comes as it matches 2nd row & 4th column of File1.txt with 4th row of File2.txt)
NEW|abcd|1234|29000000|Hello|New_value|123    (this row comes as it matches 2nd row & 4th column of File1.txt with 5th row of File2.txt)
NEW|xyzq|5678|20000000|myname|New_Value|123   (this row comes as it matches 3rd row & 4th column of File1.txt with 4th row of File2.txt)

我可以写一个像下面这样的解决方案，它也给了我正确的输出。但是当 File1.txt 和 File2.txt 都有大约 150K 行时，这需要 21 分钟。生成的最终目标文件中包含超过 1000 万行。

VAL1=123

for ROW in `cat File1.txt`
do
  Fld1=`echo $ROW | cut -d'|' -f'1-3'`
  Fld2=`echo $ROW | cut -d'|' -f4`
  Fld3=`echo $ROW | cut -d'|' -f'5-6'`

  grep -i $Fld2 File2.txt | cut -d'|' -f1  > File3.txt
  sed 's/^/'$Fld1'|/g' File3.txt | sed 's/$/|'${Fld3}'|'${VAL1}'/g' >> Target.txt

done 

但我的问题是这个解决方案可以优化吗？可以使用 AWK 或任何其他方式更快地重写它吗？

标签： unixawk

Answer 1

我很确定这会更快（因为在单个 awk 或 sed 进程中使用隐式循环通常比在 shell 循环中一遍又一遍地调用它更快），但你必须尝试它并让我们知道：

编辑：这个版本应该解决输出中重复的问题

$ cat a.awk
NR == FNR {
    for (i=1; i<=NF; ++i) {
        if ($i in a)
            a[$i] = a[$i] "," $1
        else
            a[$i] = $1;
    }
    next 
}

$4 in a {
    split(a[$4], b, ",")
    for (i in b) {
        if (!(b[i] in seen)) {
            print $1, $2, $3, b[i], $5, $6, new_value
            seen[b[i]]
        }
    }
    delete seen
}

输出包含所需的行，尽管顺序不同：

$ awk -v new_value=123 -v OFS="|" -f a.awk FS='[|>]' file2.txt FS='|' file1.txt 
NEW|abcd|1234|19000000|Hello|New_value|123
NEW|abcd|1234|17000000|Hello|New_value|123
NEW|abcd|1234|10000000|Hello|New_value|123
NEW|abcd|1234|29000000|Hello|New_value|123
NEW|abcd|1234|20000000|Hello|New_value|123
NEW|xyzq|5678|20000000|myname|New_Value|123