Question

在我的档案中，我有以下内容。

<sql type="cognos">SELECT * FROM  [PG WIND JAROS - JAROSODS].AGREMENT_MAS WHERE FILTER = 'Saravana' and ADDress = '32/36 Abc Nagar CH' </sql>
<sql type="cognos">select OBJECT_TYPE           'Object Type',
 OBJECT_NAME            "Object Name",
from [PG WIND Jaros - JAROSDM].CX_WIND_LOOKUP_REF WHERE Filter = 'Kumaar' and STATUS = 'RICH' </sql>

这里我必须将所有内容转换为小写，但单引号中的字符串除外。即，

输出：

<sql type="cognos">select * from  [pg wind jaros - jarosods].agrement_mas where filter = 'Saravana' and address = '32/36 Abc Nagar CH' </sql>
<sql type="cognos">select object_type           'Object Type',
 object_name            "object name",
from [pg wind jaros - jarosdm].cx_wind_lookup_ref where filter = 'Kumaar' and status = 'RICH' </sql>

我怎样才能在SED中实现这一目标？请帮帮我......

Answer 1

这是一个awk解决方案：

awk -F\' '{for (i=1;i<=NF;i++) printf "%s"(NF==i?RS:FS),(i%2?tolower($i):$i)}' file
<sql type="cognos">select * from  [pg wind jaros - jarosods].agrement_mas where filter = 'Saravana' and address = '32/36 Abc Nagar CH' </sql>
<sql type="cognos">select object_type           'Object Type',
 object_name            "object name",
from [pg wind jaros - jarosdm].cx_wind_lookup_ref where filter = 'Kumaar' and status = 'RICH' </sql>

Answer 2

以下是sed解决方案：

sed -e "s/^\([^']*\)$/\L\1/" -e "s/\([^']*\)\('[^']*'\)/\L\1\E\2/g" -e "s/\('[^']*\)$/\L\1/" file

\L将替换为小写（直到找到\E）。

http://www.gnu.org/software/sed/manual/sed.html

在UNIX中使用SED消除单引号字符串

2 个答案: