Conditional find / replace in bash shell

Question

My issue should be rather simple, and I'm guessing where I'm failing is syntax related. I'm looking to write a simple shell script which manipulates an XML file in the following way.

The data points inside each tag will be numeric values. These should all be positive, so if there exists a negative value inside any tag within the XML file, I would like to replace it with a zero. For example, the following XML file - call it "file.xml"

<tag1>19</tag1>
<tag2>2</tag2>
<tag3>-12</tag3>
<tag4>37</tag4>
<tag5>-41</tag5>

should be replaced with

<tag1>19</tag1>
<tag2>2</tag2>
<tag3>0</tag3>
<tag4>37</tag4>
<tag5>0</tag5>

My thinking on this would be if I grepped any instance of the string ">-*<" in the file and used sed to replace it with >0< as follows.

#!/bin/bash
STRING=">-*<"
if grep -xq "$STRING" file.xml
then
sed -i 's/$STRING/>0</g' file.xml
else
echo "that string was not found in the file"
fi

However all I'm getting in return is the echo string "that string was not found in the file" being returned, even tho I have included negative values in the file. Does the * not take into account any string following the minus sign in this example? Naturally there can be any number following the minus sign, so i'm thinking my problem is how I've defined the variable: STRING=">-*<".... Any pointers in the right direction would be greatly appreciated. Thanks in advance.

Jochen Müller · Accepted Answer · 2016-01-15 14:45:29Z

2

Try this:

STRING=">-.*<"

or better yet:

STRING=">-[0-9]*<"

In general, the * means 'any number of the last character/class of characters', so .* matches any string, [0-9]* any string consisting only of digits. Your expression would have matched '><', '>-<', '>--<', '>---<' and so on.

answered Jan 15, 2016 at 14:45

Jochen Müller

3091 silver badge6 bronze badges

Sign up to request clarification or add additional context in comments.

1 Comment

tasslebear Over a year ago

Thanks for the response @jomuel. You were spot on. The "-xq" option was also messing me up. I replaced this with just "-q" and it worked fine. I'll accept your answer now, and thanks again for the help.

bobo · Accepted Answer · 2016-01-15 15:54:17Z

2

cat aaa.txt
<tag1>19</tag1>
<tag2>2</tag2>
<tag3>-12</tag3>
<tag4>37</tag4>
<tag5>-41</tag5>

replace negatives with zero and write to a new file

sed 's/-[0-9][0-9]*/0/' aaa.txt > a2.txt

check the new file

cat a2.txt
<tag1>19</tag1>
<tag2>2</tag2>
<tag3>0</tag3>
<tag4>37</tag4>
<tag5>0</tag5>

answered Jan 15, 2016 at 15:54

bobo

931 silver badge7 bronze badges

1 Comment

tasslebear Over a year ago

Thanks for the response @bobo. Yep that definitely works alright. I have the code working fine now. Thanks again for taking the time to respond.

Calvin Taylor · Accepted Answer · 2016-01-15 15:55:24Z

1

>-[0-9]+<

is a better choice as it is one or more chars

answered Jan 15, 2016 at 15:55

Calvin Taylor

7044 silver badges17 bronze badges

3 Comments

tasslebear Over a year ago

Is there any instance where >-.*< might fail?

Calvin Taylor Over a year ago

yes, if you have any other xml included in your file that has text starting with dashes.

tasslebear Over a year ago

Of course. In this particular instance there will only ever be numeric values in the xml file so I should be safe, but point duly noted. Thanks

Collectives™ on Stack Overflow

Conditional find / replace in bash shell

3 Answers 3

1 Comment

1 Comment

3 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

3 Answers 3

1 Comment

1 Comment

3 Comments

Your Answer

Sign up or log in

Post as a guest

Related