BASH: Split strings without any delimiter and keep only first sub-string - regex

Question

I have a CSV file containing 7 columns and I am interested in modifying only the first column. In fact, in some of the rows a row name appears n times in a concatenated way without any space. I need a script that can identify where the duplication starts and remove all duplications.
Example of a row name among others:
Row name = EXAMPLE1.ABC_DEF.panel4EXAMPLE1.ABC_DEF.panel4EXAMPLE1.ABC_DEF.panel4
Replace by: EXAMPLE1.ABC_DEF.panel4
In the different rows:
n can vary
The length of the row name can vary
The structure of the row name can vary (eg. amount of _ and .), but it is always collated without any space
What I have tried:
:%s/(.+)\1+/\1/
Step-by-step:
%s: substitute in the whole file
(.+)\1+: First capturing group. .+ matches any character (except for line terminators), + is the quantifier — matches between one and unlimited times, as many times as possible, giving back as needed.
\1+: matches the same text as most recently matched by the 1st capturing group
Substitute by \1
However, I get the following errors:
E65: Illegal back reference
E476: Invalid command

Answer 1

From what i understand you need only one line contain EXAMPLE1.ABC_DEF.panel4. In that case you can do the following:
First remove duplicates in one line:
sed -i "s/EXAMPLE1.ABC_DEF.panel4.*/EXAMPLE1.ABC_DEF.panel4/g"
Then remove duplicated lines:
awk '!a[$0]++'

Answer 2

If all your rows are of the format you gave in the question (like EXAMPLExyzEXAMPLExyz) then this should work-
awk -F"EXAMPLE" '{print FS $2}' file
This takes "EXAMPLE" as the field delimiter and asks it to print only the first 'column'. It prepends "EXAMPLE" to this first column (by calling the inbuilt awk variable FS). Thanks, #andlrc.
Not an ideal solution but may be good enough for this purpose.

Answer 3

This script, with first arg is the string to test, can retrieve the biggest duplicate substring (i.e. "totototo" done "toto", not "to")
#!/usr/bin/env bash
row_name="$1"
#test duplicate from the longest to the smallest, by how many we need to split the string ?
for (( i=2; i<${#row_name}; i++ ))
do
match="True"
#continue test only if it's mathematically possible
if (( ${#row_name} % i )); then
continue
fi
#length of the potential duplicate substring
len_sub=$(( ${#row_name} / i ))
#test if the first substring is equal to each others
for (( s=1; s<i; s++ ))
do
if ! [ "${row_name:0:${len_sub}}" = "${row_name:$((len_sub * s)):${len_sub}}" ]; then
match="False"
break
fi
done
#each substring are equal, so return string without duplicate
if [ $match = "True" ]; then
row_name="${row_name:0:${len_sub}}"
break
fi
done
echo "$row_name"