根据第一列组合几个制表符分隔文件的某些列

Question

inFile 中的第一列包含一个不一定存在于所有 inFiles 中的字符串

每个 inFile 中的第 2 和第 7 列包含 Title# 字符串

使用 AWK，我无法正确地将其拼凑在一起。我对描述性变量的使用有望帮助澄清我正在尝试做的事情。这些是我认为我需要的组件：

1. 制表符分隔的输入文件：-F'\t'
2. 增加第一列中的字符串，但每个“名称”只添加一次到“1stColumnNames”：!1stColumnNames[$1]++ { name[++i] = $1 }
3. 为每个 .tsv 文件创建一个新索引来存储每个文件的值以避免覆盖每列的值：!r[FILENAME]++ { ++argind }
4. 在每个文件的第 2 列和第 7 列中存储对应的列值：{ 2ndColumnVals[$1, argind] = $2 } { 7thColumnVals[$1, argind] = $7 }
5. 打印所有 1stColumnNames 以及相关的 2ndColumnVals 和 7thColumnVals，包括它们的标题 'Title1' 'Title2' 'Title3' 等：?????
6. 对于特定 2ndColumnVals 或 7thColumnVals 为空的索引值，打印为 Mtee：?????
7. 对当前工作目录中的所有 .tsv 文件执行此操作并输出一个新的 tsv 文件：*.tsv > outFile.tsv

示例文件

inFile1.tsv

Names   Title1  Title2
AAAA    1111    123456
BBBBB   1111    123456
CCC     1111    123456

inFile2.tsv

Names   Title3  Title4
BBBBB   2222    789456
DDDDD   2222    789456
EEEE    2222    789456

inFile3.tsv

Names   Title5  Title6
AAAA    3333    987654
CCC     3333    987654
EEEE    3333    987654

outFile123.tsv

Names   Title1  Title2  Title3  Title4  Title5  Title6
AAAA    1111    123456  Mtee    Mtee    3333    987654  
BBBBB   1111    123456  2222    789456  Mtee    Mtee
CCC     1111    123456  Mtee    Mtee    3333    987654
DDDDD   Mtee    Mtee    2222    789456  Mtee    Mtee
EEEE    Mtee    Mtee    2222    789456  3333    987654

测试脚本

GNU Awk 4.0.1 位于 /usr/bin/awk ，所以我制作了这个文件并在 3 个输入文件所在的同一工作目录中执行它：

命名为 script1.sh

#### Example Usage:  script1.sh inFile1.tsv inFile2.tsv inFile3.tsv > outFile123.tsv

awk -F'\t' '
FNR==1 { ++numFiles}
!seen[$1]++ { keys[++numKeys] = $1 }
{ a[$1,numFiles] = $2 FS $3 }
END {
    for (keyNr=1; keyNr<=numKeys; keyNr++) {
        key = keys[keyNr]
        printf "%s", key
        for (fileNr=1;fileNr<=numFiles;fileNr++) {
            printf "\t%s", ((key,fileNr) in a ? a[key,fileNr] : "Mtee\tMtee")
        }
        print ""
    }
}
' "$@"

运行 awk -F script1.awk inFile1.tsv inFile2.tsv inFile3.tsv > outFile123.tsv 会打印以下错误消息：

awk: cmd. line:1: inFile1.tsv

awk: cmd. line:1: ^ syntax error

来自 konsolebox 的测试脚本 2

完美运行，但我试图通过评论来理解每一行：

#!/usr/bin/awk -f
#### named as script2.awk
#### Example Usage:  awk -f script2.awk inFile1.tsv inFile2.tsv inFile3.tsv > outFile123.tsv

BEGIN { FS = "\t" } #input File Style is tab-delimited
{ sub(/\r/, "") }   #remove all carriage return characters
!f[FILENAME]++ { ++indx }   #for all files inputted make a single index called indx
!a[$1]++ { keys[i++] = $1 } #the new indx comprises only unique strings in column 1
{ b[$1, indx] = $2 FS $3 }  #the 2nd and 3rd column are tab delimited and each pair that corresponds to a string saved in keys gets stored after the 1st column string in matrix b
END {
    for (i = 0; i in keys; ++i) {   #????
        key = keys[i]   #????
        printf "%s", keys   #prints out all strings in the index column 1 stored as keys
        for (j = 1; j <= indx; ++j) {   #????
            v = b[key, j]   #????
            printf "\t%s", length(v) ? v : "Mtee" FS "Mtee" #print out strings as tab delimited and replace any lengths of 1 char to two Mtee separated by a tab
        }
        print ""    #????
    }
}

Answer 1

你需要这样的东西：

Gawk 版本（用于 Gawk 4.0+ 中的 ARGIND 加上真正的 2D 数组）：

$ gawk -F'\t' '
!seen[$1]++ { keys[++numKeys] = $1 }
{ a[$1][ARGIND] = $2 FS $3 }
END {
    for (keyNr = 1; keyNr <= numKeys; keyNr++) {
        key = keys[keyNr]
        printf "%s", key
        for (fileNr = 1; fileNr <= ARGIND; fileNr++) {
            printf "\t%s", (fileNr in a[key] ? a[key][fileNr] : "Mtee\tMtee")
        }
        print ""
    }
}
' file1 file2 file3

非 gawk 版本：

awk -F'\t' '
FNR==1 { ++numFiles}
!seen[$1]++ { keys[++numKeys] = $1 }
{ a[$1,numFiles] = $2 FS $3 }
END {
    for (keyNr=1; keyNr<=numKeys; keyNr++) {
        key = keys[keyNr]
        printf "%s", key
        for (fileNr=1;fileNr<=numFiles;fileNr++) {
            printf "\t%s", ((key,fileNr) in a ? a[key,fileNr] : "Mtee\tMtee")
        }
        print ""
    }
}
' file1 file2 file3
Names   Title1  Title2  Title3  Title4  Title5  Title6
AAAA    1111    123456  Mtee    Mtee    3333    987654
BBBBB   1111    123456  2222    789456  Mtee    Mtee
CCC     1111    123456  Mtee    Mtee    3333    987654
DDDDD   Mtee    Mtee    2222    789456  Mtee    Mtee
EEEE    Mtee    Mtee    2222    789456  3333    987654

Answer 2

这是另一个awk：

#!/usr/bin/awk -f
# Set field separator to tab (\t)
BEGIN { FS = "\t" }
# Remove carriage return characters if file is in DOS format (CRLF)
{ sub(/\r/, "") }
# Increment indx by 1 (starts at 0) everytime a new file is processed
!f[FILENAME]++ { ++indx }
# Add a key ($1) to keys array every time it is first encountered
!a[$1]++ { keys[i++] = $1 }
# Store the 2nd and 3rd field to b matrix
{ b[$1, indx] = $2 FS $3 }
# This block runs after all files are processed
END {
    # Traverse the keys in order
    for (i = 0; i in keys; ++i) {
        key = keys[i]
        # Print key
        printf "%s", key
        # Print columns from every file in order
        for (j = 1; j <= indx; ++j) {
            v = b[key, j]
            printf "\t%s", length(v) ? v : "Mtee" FS "Mtee"
        }
        # End the line with a newline
        print ""
    }
}

用法：

awk -f script.awk file1 file2 file3

输出：

Names   Title1  Title2  Title3  Title4  Title5  Title6
AAAA    1111    123456  Mtee    Mtee    3333    987654
BBBBB   1111    123456  2222    789456  Mtee    Mtee
CCC     1111    123456  Mtee    Mtee    3333    987654
DDDDD   Mtee    Mtee    2222    789456  Mtee    Mtee
EEEE    Mtee    Mtee    2222    789456  3333    987654