与除非在双引号单卡更换空白 - 第二部分 [英] Replacing whitespace with single tab unless in double quotes - Part II

问题描述

该职位<一个href=\"http://stackoverflow.com/questions/33870465/replacing-whitespace-with-single-tab-unless-in-double-quotes\">Replacing除非在双引号中讨论的 AWK 的code单卡的空白与字符串中的所有空格双引号外面采用单制表符代替。两个答案被张贴，将正确的行为表示更换 - 除非若干或者非字母字符的组合是present输入字符串

有谁知道什么$ P $从所需更换pvents命令以及如何克服呢？

 ＆GT;猫文件
＃foo的酒吧名称产品巴兹qux＃非字母字符：无
＃foo_bar这样的名称。名称= foo的;产品=巴兹qux＃非字母字符：下划线，点，等号，分号
＃富吧+名称= foo，那么产品=巴兹qux＃非字母字符：破折号，再加上，等号，逗号，MULT。空格
 

所需的输出（当通过管道输送到猫-A ）

 ＃^ foo的伊巴尔^ ^ INAME ^ Iproduct我巴兹qux$
＃^ foo_bar这样一^ INAME = foo的;产品=巴兹qux$
＃富吧^我^ + = INAME富，产品=巴兹qux$
 

实际输出（当通过管道输送到猫-A ）

 ＆GT; AWK -v FPAT =[^] *| [^ [：空白：]] +'-v OFS ='\\ t''{$ 1 = $ 1} 1文件|猫-A
＃^ foo的伊巴尔^ ^ INAME ^ Iproduct我巴兹qux$
＃^ foo_bar这样一^ INAME = foo的;产品=巴兹^ Iqux$＃原文如此！
＃富吧^我^ + = INAME富，产品=巴兹^ Iqux$＃原文如此！＆GT; AWK'{$ 1 = $ 1} 1'= OFS'\\ T'FPAT =[^] +| [^] +'文件|猫-A
＃^ foo的伊巴尔^ ^ INAME ^ Iproduct我巴兹qux$
＃^ foo_bar这样一^ INAME = foo的;产品=巴兹^ Iqux$＃原文如此！
＃富吧^我^ + = INAME富，产品=巴兹^ Iqux$＃原文如此！
 

解决方案

它无关的特殊字符，更改为FPAT：

  FPAT ='（[^] *| [^ [：空白：]] +）+'
 

问题是，字段可能包含引号引起来的字符串和其它非空间像名称= foo的报价之外的字符;产品=巴兹qux。旧格局并没有描述这些类型的字段。

The post "Replacing whitespace with single tab unless in double quotes" discussed awk code with which all whitespaces in a string outside double quotes were replaced with single tab characters. Two answers were posted that would correctly conduct said replacement -- unless a certain number or combination of non-letter characters is present in the input strings.

Does anyone know what prevents the command from the desired replacement and how to overcome it?

> cat file
# foo bar Name product "baz qux"            # Non-letter chars: none
# foo_bar . Name=foo;product="baz qux"      # Non-letter chars: underscore, dot, equal sign, semicolon
# foo-bar +   Name=foo,product="baz   qux"  # Non-letter chars: dash, plus, equal sign, comma, mult. whitespaces

Desired output (when piped to cat -A):

# foo^Ibar^IName^Iproduct^I"baz qux"$
# foo_bar^I.^IName=foo;product="baz qux"$
# foo-bar^I+^IName=foo,product="baz   qux"$

Actual output (when piped to cat -A):

> awk -v FPAT='"[^"]*"|[^[:blank:]]+' -v OFS='\t' '{$1=$1} 1' file | cat -A
# foo^Ibar^IName^Iproduct^I"baz qux"$
# foo_bar^I.^IName=foo;product="baz^Iqux"$  # sic!
# foo-bar^I+^IName=foo,product="baz^Iqux"$  # sic!

> awk '{$1=$1}1' OFS='\t' FPAT='"[^"]+"|[^ ]+' file | cat -A
# foo^Ibar^IName^Iproduct^I"baz qux"$
# foo_bar^I.^IName=foo;product="baz^Iqux"$  # sic!
# foo-bar^I+^IName=foo,product="baz^Iqux"$  # sic!

解决方案

It has nothing to do with special characters, change FPAT to:

FPAT='("[^"]*"|[^[:blank:]"]+)+'

The problem was that fields may contain strings enclosed between quotes and other non-space characters outside of the quotes like Name=foo;product="baz qux". The old pattern didn't describe these kind of fields.

这篇关于与除非在双引号单卡更换空白 - 第二部分的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！