[SOLVED] sed edit the line above a regex expression

corcodelagaze · 02-25-2012, 05:39 AM

hello ,
for the last 2 days i`m dealing with teh following problem
i have a huge xml files and i want change the line that is above the finded regex

for example
xml file looks like
<grupo>AUDIO</grupo>
<familia>Tarjeta de Sonido Interna</familia>
<marca>CREATIVE</marca>

and it should look like
<grupo>Componentpc</grupo>
<familia>Tarjeta de Sonido Interna</familia>
<marca>CREATIVE</marca>

i did
cat 000615.xml |sed -n '/Tarjeta de Sonido Interna/{g;1!p;};h'
<grupo>AUDIO</grupo>
<grupo>AUDIO</grupo>
<grupo>AUDIO</grupo>
<grupo>AUDIO</grupo>
<grupo>AUDIO</grupo>
and extracted the line above but now i`m stucked `cause i don`t know how to change the line
help!

boss is killing me

druuna · 02-25-2012, 05:48 AM

Hi,

Assuming the given example is correct, here's one way of doing just that:

Code:

grep -B2 "<marca>CREATIVE</marca>" 000615.xml | sed '/>AUDIO</s/AUDIO/Componentpc/'

Test run:

Code:

$ cat 000615.xml
<grupo>AUDIO</grupo>
<familia>Tarjeta de Sonido Interna</familia>
<marca>CREATIVE</marca>
<grupo>AUDIO</grupo>
<familia>Tarjeta de Sonido Interna</familia>
<marca>CREATIVE</marca>
<grupo>AUDIO</grupo>
<familia>Tarjeta de Sonido Interna</familia>
<marca>CREATIVE</marca>

$ grep -B2 "<marca>CREATIVE</marca>" infile | sed '/>AUDIO</s/AUDIO/Componentpc/'
<grupo>Componentpc</grupo>
<familia>Tarjeta de Sonido Interna</familia>
<marca>CREATIVE</marca>
<grupo>Componentpc</grupo>
<familia>Tarjeta de Sonido Interna</familia>
<marca>CREATIVE</marca>
<grupo>Componentpc</grupo>
<familia>Tarjeta de Sonido Interna</familia>
<marca>CREATIVE</marca>

Hope this helps.

corcodelagaze · 02-25-2012, 05:57 AM

the thing is that the file has thousands of lines

i tryed to raise the lines like this
grep -B323232332 "<familia>Tarjeta de Sonido Interna</familia>" 000615.xml | sed '/>AUDIO</s/AUDIO/Componentpc/' > aa.xml

but i see that it doesn`t do the work
cat 000615.xml |wc -l
75219
[root@ventadecoche script]# cat aa.xml |wc -l
30350

druuna · 02-25-2012, 06:01 AM

Hi,

Quote:

Originally Posted by corcodelagaze

the thing is that the file has thousands of lines

i tryed to raise the lines like this
grep -B323232332 "<familia>Tarjeta de Sonido Interna</familia>" 000615.xml | sed '/>AUDIO</s/AUDIO/Componentpc/' > aa.xml

but i see that it doesn`t do the work
cat 000615.xml |wc -l
75219
[root@ventadecoche script]# cat aa.xml |wc -l
30350

I'm not sure what the problem is..... The solution given works on small and big files.

Grep's -B2 option tells grep to also show the 2 lines before the line a hit is matched. The sed part checks those 3 lines and only modifies the line that holds AUDIO.

Please elaborate about the problem you are having.

corcodelagaze · 02-25-2012, 06:10 AM

first of all thank you for your help !!!
i have a xml file whith thousands of lines and i need to change some lines
000615.xml looks like
<producto num="1267">
<codigo>CR AUDIGY SE OEM</codigo>
<codigo_2>-CR AUDIGY SE OEM</codigo_2>
<codigo_3>SB0570</codigo_3>
<codigo_4 />
<grupo>AUDIO</grupo>
<familia>Tarjeta de Sonido Interna</familia>
<marca>CREATIVE</marca>
<precio>14.68</precio>
<stock>68</stock>
<peso>0.00</peso>
<nombre>CREATIVE TARJETA DE SONIDO AUDIGY SE OEM PCI</nombre>
<descripcion />
<caracteristicas />
<imagen>http://www.sss.com/1.jpg</imagen>
</producto>
i need to change <grupo>AUDIO</grupo> into <grupo>componentspc</grupo> only where it finds <familia>Tarjeta de Sonido Interna</familia>
the logic scheme would look like
cat 000615.xml | find the lines <familia>Tarjeta de Sonido Interna</familia> and change the line obove it from <grupo>AUDIO</grupo> into

<grupo>componentspc</grupo>

corcodelagaze · 02-25-2012, 06:18 AM

uhh oh i need to replace the line <grupo>AUDIO</grupo> into <grupo>Componentpc</grupo> in the file 000615.xml

druuna · 02-25-2012, 06:23 AM

Hi,

Give this a try:

Code:

sed '/AUDIO/{N;/Tarjeta de Sonido Interna/s/AUDIO/Componentspc/}' 000615.xml

This works on the small example you have given. It does assume that the line containing Tarjeta de Sonido Interna is always below the line containing AUDIO.

Hope this helps.

corcodelagaze · 02-25-2012, 06:27 AM

yupyy !!! that dit the job thank thank you !!!

druuna · 02-25-2012, 06:29 AM

Quote:

Originally Posted by corcodelagaze

yupyy !!! that dit the job thank thank you !!!

You're welcome

BTW: Can you put up the [SOLVED] tag.
first post -> Thread Tools -> Mark this thread as solved

ntubski · 02-25-2012, 08:52 AM

Alternate solution with xmlstarlet:

Code:

xml ed -u '//producto[familia="Tarjeta de Sonido Interna"]/grupo[.="AUDIO"]' -v componentspc 000615.xml