tidyverse概念的awk等价物（熔化和扩散）

我有一些文本日志需要解析并格式化为CSV。 我有一个可以工作的R脚本，但一旦文件大小增加，它就会变慢，据我所知，这个问题似乎是使用awk（或其他命令行工具？）加速的一个很好的候选

我在

awk

方面做得不多，我面临的问题是如何将我对R中处理的想法转换为

awk

脚本编写的方式

示例截断的输入数据（Scrap.log）： 预期输出（截断）： 我在R脚本中的一般步骤：

在文件顶部添加具有新名称的单个标题行

每行的最上面一行（以！！G开头）

标题列（_START）的格式从宽到长

到目前为止，我在awk工作的作品包括：

如何抓取和打印标题行

awk'/\u START/{header=$0；print header}'废料.log

如何使用新的标题值写入单行

awk'BEGIN{ORS=”“；for（counter=1；counter我认为：

awk '
BEGIN{
    # output the header line
    print "HH1,HH2,HH3,HH4,HH5,HH6,HH7,HH8,HH9,HH10,HH11,HH12,HH13,HH14,START,HeaderName,Value"
}
# ignore comment lines
/;/{next}

/!!G/{
    valcnt = 1
    # save and shuffle the values
    val[valcnt++] = $2
    val[valcnt++] = $11
    val[valcnt++] = $12
    val[valcnt++] = $13
    val[valcnt++] = $14
    val[valcnt++] = $15
    val[valcnt++] = $3
    val[valcnt++] = $4
    val[valcnt++] = $5
    val[valcnt++] = $6
    val[valcnt++] = $7
    val[valcnt++] = $8
    val[valcnt++] = $9
    val[valcnt++] = $10
    next
}
/_START /{
    # these are headers - save them to head, to be reused later
    for (i = 2; i <= NF; ++i) {
        # fun fact: its indexed on NF
        head[i] = $i
    }
    next
}

# this function is redundant, but its just easier for me to think about the code
function output(firstval, header, value, \
        cur, i) {
    cur = valcnt
    val[cur++] = firstval
    val[cur++] = header
    val[cur++] = value
    # output val as csv
    for (i = 1; i < cur; ++i) {
        printf "%s%s", val[i], i != cur - 1 ? "," : "\n"
    }
}

/[0-9]+/{
    for (i = 2; i <= NF; ++i) {
        # add these 3 to all the other values and output them
        # ie. add first column, the header from header and the value
        output($1, head[i], $i)
    }
}

'

awk'
开始{
#输出标题行
打印“HH1、HH2、HH3、HH4、HH5、HH6、HH7、HH8、HH9、HH10、HH11、HH12、HH13、HH14、开始、标题名称、值”
}
#忽略注释行
//{next}
/！！G/{
valcnt=1
#保存并洗牌这些值
val[valcnt++]=$2
val[valcnt++]=$11
val[valcnt++]=12美元
val[valcnt++]=$13
val[valcnt++]=$14
val[valcnt++]=$15
val[valcnt++]=$3
val[valcnt++]=$4
val[valcnt++]=$5
val[valcnt++]=$6
val[valcnt++]=$7
val[valcnt++]=$8
val[valcnt++]=9美元
val[valcnt++]=10美元
下一个
}
/_开始/{
#这些是标题-将它们保存到标题，以便以后重用
对于（i=2；如上所述，它在
output
中的循环末尾过早跳行，如：
99999，CURRENT，XYZ，ABC，STATE1，STATE2，%，%，%，%，%，%，%，%，%，10，Header12,44.28
99999，CURRENT，XYZ，ABC，STATE1，STATE2，%，%，%，%，%，%，STATE1，Header13,16.57
99999，CURRENT，XYZ，ABC，STATE1，STATE1，STATE1，%，%，%，%，Header1，STATE2，%，%，%，%，Header1，%，%4
，11.52
我可以通过将最后一个函数中的等式更改为：
/[0-9]+/{for（I=2；I
它将删除一行，您知道吗？请确保没有dos行结尾。啊，您是正确的。该更改将删除两个块中的
头14
行。我必须找到另一种方法来防止行跳过
HH1,HH2,HH3,HH4,HH5,HH6,HH7,HH8,HH9,HH10,HH11,HH12,HH13,HH14,START,HeaderName,Value
99999,CURRENT,XYZ,ABC,STATE1,STATE2,%,%,%,%,%,%,%,%,10,Header1,12.23
99999,CURRENT,XYZ,ABC,STATE1,STATE2,%,%,%,%,%,%,%,%,10,Header2,1.91
99999,CURRENT,XYZ,ABC,STATE1,STATE2,%,%,%,%,%,%,%,%,10,Header3,6.63
99999,CURRENT,XYZ,ABC,STATE1,STATE2,%,%,%,%,%,%,%,%,10,Header4,1.68
99999,CURRENT,XYZ,ABC,STATE1,STATE2,%,%,%,%,%,%,%,%,10,Header5,50.03
99999,CURRENT,XYZ,ABC,STATE1,STATE2,%,%,%,%,%,%,%,%,10,Header6,0.5
99999,CURRENT,XYZ,ABC,STATE1,STATE2,%,%,%,%,%,%,%,%,10,Header7,13.97
99999,CURRENT,XYZ,ABC,STATE1,STATE2,%,%,%,%,%,%,%,%,11,Header1,11.32
99999,CURRENT,XYZ,ABC,STATE1,STATE2,%,%,%,%,%,%,%,%,11,Header2,1.94
99999,CURRENT,XYZ,ABC,STATE1,STATE2,%,%,%,%,%,%,%,%,11,Header3,6.64
99999,CURRENT,XYZ,ABC,STATE1,STATE2,%,%,%,%,%,%,%,%,11,Header4,1.94
99999,CURRENT,XYZ,ABC,STATE1,STATE2,%,%,%,%,%,%,%,%,11,Header5,50.12
99999,CURRENT,XYZ,ABC,STATE1,STATE2,%,%,%,%,%,%,%,%,11,Header6,0.58
99999,CURRENT,XYZ,ABC,STATE1,STATE2,%,%,%,%,%,%,%,%,11,Header7,15.1
99999,CURRENT,XYZ,ABC,STATE1,STATE2,%,%,%,%,%,%,%,%,12,Header1,12.96
99999,CURRENT,XYZ,ABC,STATE1,STATE2,%,%,%,%,%,%,%,%,12,Header2,2.15
99999,CURRENT,XYZ,ABC,STATE1,STATE2,%,%,%,%,%,%,%,%,12,Header3,6.57
99999,CURRENT,XYZ,ABC,STATE1,STATE2,%,%,%,%,%,%,%,%,12,Header4,2.12
99999,CURRENT,XYZ,ABC,STATE1,STATE2,%,%,%,%,%,%,%,%,12,Header5,55.6
99999,CURRENT,XYZ,ABC,STATE1,STATE2,%,%,%,%,%,%,%,%,12,Header6,0.62
99999,CURRENT,XYZ,ABC,STATE1,STATE2,%,%,%,%,%,%,%,%,12,Header7,16.24
99999,CURRENT,XYZ,ABC,STATE1,STATE2,%,%,%,%,%,%,%,%,13,Header1,11.43
99999,CURRENT,XYZ,ABC,STATE1,STATE2,%,%,%,%,%,%,%,%,13,Header2,2.18
...


awk '
BEGIN{
    # output the header line
    print "HH1,HH2,HH3,HH4,HH5,HH6,HH7,HH8,HH9,HH10,HH11,HH12,HH13,HH14,START,HeaderName,Value"
}
# ignore comment lines
/;/{next}

/!!G/{
    valcnt = 1
    # save and shuffle the values
    val[valcnt++] = $2
    val[valcnt++] = $11
    val[valcnt++] = $12
    val[valcnt++] = $13
    val[valcnt++] = $14
    val[valcnt++] = $15
    val[valcnt++] = $3
    val[valcnt++] = $4
    val[valcnt++] = $5
    val[valcnt++] = $6
    val[valcnt++] = $7
    val[valcnt++] = $8
    val[valcnt++] = $9
    val[valcnt++] = $10
    next
}
/_START /{
    # these are headers - save them to head, to be reused later
    for (i = 2; i <= NF; ++i) {
        # fun fact: its indexed on NF
        head[i] = $i
    }
    next
}

# this function is redundant, but its just easier for me to think about the code
function output(firstval, header, value, \
        cur, i) {
    cur = valcnt
    val[cur++] = firstval
    val[cur++] = header
    val[cur++] = value
    # output val as csv
    for (i = 1; i < cur; ++i) {
        printf "%s%s", val[i], i != cur - 1 ? "," : "\n"
    }
}

/[0-9]+/{
    for (i = 2; i <= NF; ++i) {
        # add these 3 to all the other values and output them
        # ie. add first column, the header from header and the value
        output($1, head[i], $i)
    }
}

'