I have a .prog script which is filtering ASCII characters and special characters from a XML output file.

Now, in a new requirement, Cyrillic characters/language will also be printed in the XML output. I want that my script should allow them also to get printed.

Currently, I am using below SED command:

sed -e "2,$ s/[^a-zA-Z0-9/?).,'+{}\n\r<>_"= -]//g" $cm_file > $work_file

Can anyone please help me in this?

Regards,
Gurpreet

Using sed to parse xml files is pretty much always going to be a nightmare at some point. Your script seems to be looking for data to remove rather than what to keep. As your list of what to keep is
growing, I would suggest changing it to select what to remove, or better yet use a language like perl, python , ruby, etc which can parse xml correctly and allow you to display what you want / need.

2 members found this post helpful.

Hi,

Yes, you will need a higher level programming language, like Perl or Python. sed is very powerful but the character set is ASCII. You can tell your system to use UTF-8 characters by changing the locale but you will need to set up a "dictionary" of these characters in order for a program to do comparison. This sounds more like a job in Python.

And as grail said, the task might be easier if you choose to select what to remove instead.

I like your suggestion, but we have a limitation here. I can not use any other language. If I go for selecting what to remove, then can you please guide how to do that since I am not very comfortable with SED command.

You will need to supply an example of what you are starting with and what the final output should be?

Also, please use the [code][/code] tags around your examples to maintain formatting and help with readability

Hi,

We will have a XML output file which will contain the Cyrillic language characters as well apart from the allowable characters as mentioned in the below command:

sed -e "2,$ s/[^a-zA-Z0-9/?).,'+{}\n\r<>_"= -]//g" $cm_file > $work_file

In the XML file, we have a tag <Name> </Name>. Currently, it is passing the ASCII characters only ,

like for eg. <Name>ABCD109</Name.

Now, the requirement says Cyrillic characters/language will come this tag,

<Name>Обозначения использования</Name>

Initially, I thought like ASCII characters, I would add the Cyrillic characters range, but that is not possible.

Can you please share how to modify this sed command or create a new one so that this tag also displays the Cyrillic characters? Please help.


Regards,
Gurpreet Kaur

So, here are the several problems I have with the last post:

1. No use of 'code' tags as requested

2. No before and after example

3. As mentioned, sed is the wrong tool for non-ASCII related work

New information:

1. sed is also the wrong tool to use to parse XML

So, once again, please provide an actual example or be more clear about the example you have provided?
Using the above as a possible example, what exactly is wrong with the data shown? What exactly are you trying to remove / keep?

As mentioned by aragorn2101, perl, python or ruby would be the better choices to parse XML data.