[文本处理] [已解决]怎样突破批处理 For /F 的 tokens 能处理的令牌数限制？

hnfeng

有一 csv 文件，以“,” 分割，共有47列
想导出这些列 1,14,26,38,40,42 重新生成一个 csv 文件，但是貌似 for /f "skip=1 tokens=1,14,26,38,40,42 delims=," %%a in (test.csv) do ………… 是不成功的，因为 tokens 好像最大只能处理31个令牌。

目前的思路，是先获取 1,14 列和 26列之后的，然后再处理 获得的 26列以后的字符……，感觉蛮麻烦的。找了找 第三方库，也没找到合适的第三方程序。

请高手帮忙指教，谢谢

这是前几行的，后面还有一些：

,/lpc/it8620e/control/0,/lpc/it8620e/control/1,/lpc/it8620e/control/2,/lpc/it8620e/voltage/0,/lpc/it8620e/voltage/1,/lpc/it8620e/voltage/2,/lpc/it8620e/voltage/3,/lpc/it8620e/voltage/4,/lpc/it8620e/voltage/5,/lpc/it8620e/voltage/6,/lpc/it8620e/voltage/7,/lpc/it8620e/voltage/8,/lpc/it8620e/temperature/0,/lpc/it8620e/temperature/2,/lpc/it8620e/fan/0,/intelcpu/0/load/1,/intelcpu/0/load/2,/intelcpu/0/load/3,/intelcpu/0/load/4,/intelcpu/0/load/0,/intelcpu/0/temperature/0,/intelcpu/0/temperature/1,/intelcpu/0/temperature/2,/intelcpu/0/temperature/3,/intelcpu/0/temperature/4,/intelcpu/0/clock/1,/intelcpu/0/clock/2,/intelcpu/0/clock/3,/intelcpu/0/clock/4,/intelcpu/0/power/0,/intelcpu/0/power/1,/intelcpu/0/power/2,/intelcpu/0/power/3,/intelcpu/0/clock/0,/hdd/0/temperature/0,/hdd/0/load/0,/hdd/1/temperature/0,/hdd/1/load/0,/hdd/2/temperature/0,/hdd/2/load/0,/hdd/3/temperature/0,/hdd/3/load/0,/hdd/4/temperature/0,/hdd/4/load/0,/hdd/5/temperature/0,/hdd/5/load/0
Time,Fan Control #1,Fan Control #2,Fan Control #3,Voltage #1,Voltage #2,Voltage #3,Voltage #4,Voltage #5,Voltage #6,Voltage #7,Standby +3.3V,VBat,Temperature #1,Temperature #3,Fan #1,CPU Core #1,CPU Core #2,CPU Core #3,CPU Core #4,CPU Total,CPU Core #1,CPU Core #2,CPU Core #3,CPU Core #4,CPU Package,CPU Core #1,CPU Core #2,CPU Core #3,CPU Core #4,CPU Package,CPU Cores,CPU Graphics,CPU DRAM,Bus Speed,Temperature,Used Space,Temperature,Used Space,Temperature,Used Space,Temperature,Used Space,Temperature,Used Space,Temperature,Used Space
05/24/2024 00:03:59,,,,0.156,2.04,2.052,2.04,0.012,1.728,1.512,3.384,3.048,28,23,1912.18127,7.69230747,41.53846,16.9230766,6.153846,18.07692,31,28,21,26,31,800.0051,800.0051,800.0051,800.0051,7.07779026,0.8458846,0.002768853,2.3195765,100.000641,27,28.538063,42,97.98176,39,49.0412674,43,97.9916458,30,88.13136,32,90.882576
05/24/2024 00:08:59,,,,0.420000017,2.04,2.052,2.04,0.012,1.728,1.512,3.384,3.048,28,23,1912.18127,32.30769,7.69230747,16.9230766,3.076923,14.9999981,31,29,21,25,31,800.004761,800.004761,800.004761,800.004761,6.94404268,0.880736053,0.002829046,2.232779,100.000595,27,28.5380554,42,97.98176,39,49.0412674,43,97.9916458,30,88.13136,32,90.882576
05/24/2024 00:13:59,,,,0.744,2.04,2.052,2.04,0.012,1.728,1.512,3.384,3.048,28,23,1917.61365,16.9230766,0,0,13.8461533,7.69230747,32,28,20,25,31,800.005,800.005,800.005,800.005,5.744312,0.430073529,0.00227422,2.047413,100.000626,27,28.5380554,42,97.98176,40,49.0412674,43,97.9916458,30,88.13136,32,90.882576
复制代码

newswan

回复 1# hnfeng


  用 awk

 awk -F',' 'BEGIN{OFS=FS} {print $1,$14,$26,$38,$40,$42}' %filename%
复制代码

hnfeng

回复  hnfeng


  用 awk
newswan 发表于 2024-5-25 15:20

   请问哪里下载 awk ？ http://www.bathome.net/s/tool/?key=awk 这里找到的 gawk，按上面的语句出错

newswan

回复 3# hnfeng


msys 这个比较大，你试试
https://www.msys2.org/

gawk64.exe
单引号 改 双引号

gawk64.exe -F"," "BEGIN{OFS=FS} {print $1,$14,$26,$38,$40,$42}" "%filename%"
复制代码

娜美

cut -d , -f 1-32 file
复制代码

aloha20200628

回复 1# hnfeng

给一个纯P版本，以下代码存为test.bat运行，提取结果文件存为 test.new.csv
假设源文件*.csv是ansi(即简中)编码，每行数据中未包含制表符和等号
提取列数可在第2行自定义，只须用斜线符包裹每一个列数即可...

@echo off &setlocal enabledelayedexpansion
set "lst=/1/14/26/38/40/42/"
(for /f "delims=" %%a in (test.csv) do (
	set "a=%%~a"&set "a=!a: =┴!"&set "a=!a:;=┬!"&set "n=0"&set "v="
	for %%b in (!a!) do (
		set/a "n+=1"
		for %%i in (!n!) do if "!lst:/%%i/=!" neq "!lst!" (set "v=!v!,%%b")
	)
	set "v=!v:┴= !"&set "v=!v:┬=;!"&echo,!v:~1!
))>"test.new.csv"
endlocal&pause&exit/b
复制代码

hnfeng

回复  hnfeng


gawk64.exe -F"," "BEGIN{OFS=FS} {print $1,$14,$26,$38,$40,$42}" "%filename%"

这个命令完美。多谢多谢

hnfeng

娜美 发表于 2024-5-25 16:15

    下载了 cut.exe，试了一下，没看到结果。可能因为我不会用吧

hnfeng

回复  hnfeng

给一个纯P版本，以下代码存为test.bat运行，提取结果文件存为 test.new.csv
假设源文件* ...
aloha20200628 发表于 2024-5-25 17:08

谢谢。但是结果不对。
正确的应该是

Time,Temperature #1,CPU Package,Temperature,Temperature,Temperature
05/24/2024 00:03:59,28,31,42,39,43
05/24/2024 00:08:59,28,31,42,39,43
05/24/2024 00:13:59,28,31,42,40,43
复制代码

批处理运行的结果是

Time,Temperature #1,CPU Package,Temperature,Temperature,Temperature
05/24/2024 00:03:59,7.69230747,800.0051,49.0412674,97.9916458,88.13136
05/24/2024 00:08:59,32.30769,800.004761,49.0412674,97.9916458,88.13136
05/24/2024 00:13:59,16.9230766,800.005,49.0412674,97.9916458,88.13136
复制代码

娜美

下载了 cut.exe，试了一下，没看到结果。可能因为我不会用吧
hnfeng 发表于 2024-5-25 19:13

cut -d , -f 1,14,26,38,40,42 file
复制代码

aloha20200628

回复 9# hnfeng

复查了一楼示例文件，其中第3-5行前端出现 ,,,, 连续空值分割符，导致运行时for循环分割器自动跳过，结果出现错位... 再给6楼代码打一个补丁吧，新代码如下，其运行结果能与楼主期望完全相符

@echo off &setlocal enabledelayedexpansion
set "lst=/1/14/26/38/40/42/"
(for /f "delims=" %%a in (test.csv) do (
	set "a=◆%%~a"&set "a=!a:,=,◆!"&set "a=!a: =┴!"&set "a=!a:;=┬!"&set "n=0"&set "v="
	for %%b in (!a!) do (
		set/a "n+=1"
		for %%i in (!n!) do if "!lst:/%%i/=!" neq "!lst!" (set "v=!v!,%%b")
	)
	set "v=!v:◆=!"&set "v=!v:┴= !"&set "v=!v:┬=;!"&echo,!v:~1!
))>"test.new.csv"
endlocal&pause&exit/b
复制代码

hnfeng

娜美 发表于 2024-5-25 20:26

    多谢多谢，此次完美。

hnfeng

回复  hnfeng

复查了一楼示例文件，其中第3-5行前端出现 ,,,, 连续空值分割符，导致运行时for循环分割器 ...
aloha20200628 发表于 2024-5-25 20:44

    多谢多谢，此次完美了。

多谢上面几位的帮忙

Five66

@echo off
(
for /f "skip=1 eol=□ delims=" %%a in (test.csv) do (
set "@line=%%a"
setlocal enabledelayedexpansion
for /f "tokens=1,14,26* delims=," %%b in ("!@line:,=□,!") do (
endlocal
set "@out=%%b,%%c,%%d"
for /f "tokens=12,14,16 delims=," %%f in ("%%e") do (
set "@out2=%%f,%%g,%%h
)
)
setlocal enabledelayedexpansion
set "str=!@out:□=!,!@out2:□=!"
echo(!str!|| ^)?
endlocal
))>the_new_test.csv
echo,done
pause&exit/b
复制代码

77七

楼主问题已解决。
将1楼内容作为文本 1.txt （不处理逗号了）。

@echo off
cd /d "%~dp0"
for /f "useback delims=" %%x in ("1.txt") do (
	setlocal
	call :getTokens "%%x" "," "1 14 26 38 40 42"
	setlocal enabledelayedexpansion
	echo=[!`str!]
	endlocal
	endlocal
)
pause
exit

	:getTokens <字符串> <分隔符> <列>
	if "%~2" neq "" (
		set "`d=%~2"
		for %%a in (%~3) do (
			set "`%%a=1"
		)
	)
	set /a tk+=1
	for /f "tokens=1* delims=%`d%" %%a in ("%~1") do (
		if defined `%tk% (
			if not defined `str (
				set "`str=%%a"
			) else (
				rem 结果以","分隔
				set "`str=%`str%,%%a"
			)
		)
		if "%%b" neq "" (
			call :getTokens "%%b"
		) else (
			exit /b
		)
	)
	exit /b
复制代码