sed无法从文件中删除

Question

I am using ubuntu shell; I am suppose to read user input and delete the line in the file. However when I run it, the terminal shows all data after deleting but my file remains untouched at all with no deletion. How do I ignore case for sed?

Data:

The Small:Sir King:23.0:30:5
The Big:Sir King:28.0:40:7
The Middle:Sir King:98.9:20:8

Code:

echo "Please enter data1: "
read data1
echo "Please enter data2: "
read data2
sed "/$data1:$data2/d" sampleDB.txt

When i type The Middle, Sir King it display this in terminal:

The Small:Sir King:23.0:30:5
The Big:Sir King:28.0:40:7

But my file still have the The Middle:Sir King:98.9:20:8

Answer 1

Use -i option to do sed replace inplace in the file as below:

sed -i "/$data1:$data2/d" sampleDB.txt

if you wish to create back up of original file before modifying the file then use it as:

sed -i.bak "/$data1:$data2/d" sampleDB.txt

Now you will have additional file sampleDB.txt.bak without any modifications and original file sampleDB.txt with lines deleted.

Answer 2

What you have now is a very dangerous way to do it. You're using regexps when you want strings, and you're matching across the whole line not specific fields.

BACK UP YOUR FILE then try it if the user inputs .* for both inputs. Now try it if they enter King and 2 . I'd guess those would NOT be desirable results.

You should really be using awk since it can operate on specific fields using string comparison and so solve both problems:

awk -F: -v d1="$data1" -v d2="$data2" '!($1==d1 && $2==d2)' sampleDB.txt > tmp &&
mv tmp sampleDB.txt

to ignore case would be:

awk -F: -v d1="$data1" -v d2="$data2" '!(tolower($1)==tolower(d1) && tolower($2)==tolower(d2))' sampleDB.txt > tmp &&
mv tmp sampleDB.txt

With GNU awk you could use IGNORECASE and -i inplace to make it briefer if you like, see the man page.

Answer 3

instead of

sed "/$data1:$data2/d" sampleDB.txt

try

sed -i "/$data1:$data2/Id" sampleDB.txt

the I flag in the command line will ignore any upper/lower case.

Answer 4

你应该试试：

sed "/d${data1}:${data2}" sampleDB.txt