如何转置gnuplot可读文件中的数据块（表格）

数据块规则：每个数据块以# description行开头，以两个空行结尾。

输入示例 #

# block_01
Temp    ST_01       ST_02       ST_03       ST_04
ag7     -0.0295167  0.0553285   0.00883727  0.392553
bc9     -0.0513702  0.0355596   -0.0165404  0.252816
cc9     -0.0303508  -0.00354388 0.0428582   0.0882037
cd3     -0.0239757  -0.0363891  0.0227775   0.270722


# block_02
Temp    ST_01       ST_02       ST_03       ST_04
ag7.b   -0.00284193 0.0260536   0.163357    -0.17291
bc9.b   -0.0212977  0.0191269   0.0330521   -0.17346  
cc9.b   -0.0138837  -0.0220729  0.0814398   -0.0478094
cd3.f   -0.053152   -0.0041982  -0.0948933  -0.020557


# block_03
Temp    ST_01       ST_02       ST_03       ST_04
...

期望输出 #

# block_01
Temp    ag7         bc9         cc9         cd3
ST_01   -0.0295167  -0.0513702  -0.0303508  -0.0239757
ST_02   0.0553285   0.0355596   -0.00354388 -0.0363891
ST_03   0.00883727  -0.0165404  0.0428582   0.0227775
ST_04   0.392553    0.252816    0.0882037   0.270722


# block_02
Temp    ag7.b       bc9.b       cc9.b       cd3.f  
ST_01   -0.00284193 -0.0212977  -0.0138837  -0.053152  
ST_02   0.0260536   0.0191269   -0.0220729  -0.0041982  
ST_03   0.163357    0.0330521   0.0814398   -0.0948933  
ST_04   -0.17291    -0.17346    -0.0478094  -0.020557


# block_03
Temp    ag7.b       bc9.b       cc9.b       cd3.f  
...

解决方案 #

1. 使用Awk脚本 #

这个脚本会逐个处理每个数据块，收集字段后转置输出：

#!/usr/bin/awk -f

/^#/ {
    if (block_lines > 0) {
        print_block()
    }
    print $0
    block_lines = 0
    next
}

/^$/ {
    empty_lines++
    if (empty_lines == 2) {
        print_block()
        empty_lines = 0
        print "\n"
    }
    next
}

{
    empty_lines = 0
    for (i=1; i<=NF; i++) {
        fields[block_lines+1][i] = $i
    }
    block_lines++
}

END {
    if (block_lines > 0) {
        print_block()
    }
}

function print_block() {
    if (block_lines == 0) return
    cols = length(fields[1])
    for (c=1; c<=cols; c++) {
        line = fields[1][c]
        for (r=2; r<=block_lines; r++) {
            line = line "\t" fields[r][c]
        }
        print line
    }
    delete fields
}

使用步骤：

1. 保存脚本为transpose_blocks.awk
2. 赋予执行权限：chmod +x transpose_blocks.awk
3. 运行处理文件：./transpose_blocks.awk input.txt > output.txt

2. 使用Bash结合GNU工具 #

通过拆分块、单独转置再合并的方式实现：

#!/bin/bash

input_file="$1"
temp_dir=$(mktemp -d)

# 拆分输入文件为单个块的临时文件
awk '
    /^#/ {
        if (outfile != "") close(outfile)
        block_num++
        outfile = sprintf("%s/block_%02d", temp_dir, block_num)
        print > outfile
        next
    }
    /^$/ {
        empty_lines++
        if (empty_lines == 2) {
            close(outfile)
            empty_lines = 0
            outfile = ""
        } else if (outfile != "") {
            print > outfile
        }
        next
    }
    {
        empty_lines = 0
        if (outfile != "") print > outfile
    }
' "$input_file"

# 处理每个临时块并输出
for block in "$temp_dir"/block_*; do
    head -n1 "$block"
    tail -n+2 "$block" | awk '{for(i=1;i<=NF;i++)a[i]=a[i]"\t"$i}END{for(i in a)print substr(a[i],2)}'
    echo -e "\n"
done | sed '$ s/\n$//'

rm -rf "$temp_dir"

使用步骤：

1. 保存脚本为transpose_blocks.sh
2. 赋予执行权限：chmod +x transpose_blocks.sh
3. 运行处理文件：./transpose_blocks.sh input.txt > output.txt

内容的提问来源于stack exchange，提问作者magfan