How To Convert Fixed Width File To Delimited Using SED In Unix/Linux

In this article, we will see how to convert a fixed width file into a delimited file using sed command. We will see step by step how sed perform this operation. Please go through with the following post How To Use & In Sed Command In Unix/Linux , if you have not.

We have a file F_Input_File.txt which have fixed width data, we have to allocate this data across 5 columns:

Col1: From 1 to 4

Col2: From 5 to 6

Col3: From 7 to 8

Col4: From 9 to 13

Col5: From 14 to 16

Input File: F_Input_File.txt

1001A1HR50000USA

Solution:

$ sed -e 's/./&,/4' -e 's/./&,/7' -e 's/./&,/10' -e 's/./&,/16' F_Inout_File.txt

Output:

1001,A1,HR,50000,USA

What is that? What exactly has happened? Let’s understand it.

Explanation: 

1. We can execute multiple sed command using -e switch.

2. dot (.) matches any number of character.

1st -e switch will read "1001A1HR50000USA" as input:

$ sed -e 's/./&,/4'

What is happening here ??

1.   Dot (.) matches everything until 4th occurrence of any character.

If you remember the syntax:
sed 's/reg_exp/replacement/[occurrence]' F_Input_File    

Occurrence	1	2	3	4
Data	1	0	0	1
REG_EXP	.	.	.	.

     2.  & will hold 1001

          Result of REG_EXP (.) = 1001

          Replacement (&,)=1001,
          After the 2nd -e switch record will be: 1001,A1HR50000USA

2nd -e switch will receive input as "1001,A1HR50000USA"

$ sed -e 's/./&,/7'

1.   dot (.) will match until 7th occurrence of any character.

Occurrence	1	2	3	4	5	6	7
Data	1	0	0	1	,	A	1
REG_EXP	.	.	.	.	.	.	.

2.   & will hold 1001,A1

Result of REG_EXP (.) = 1001,A1

Replacement (&,) =1001,A1,
After the 2nd -e switch record will be: 1001,A1,HR50000USA

3rd -e switch will receive input as "1001,A1,HR50000USA"

$ sed -e 's/./&,/10'

1.   Again, (dot) will match until 10th occurrence of any character.

Occurrence	1	2	3	4	5	6	7	8	9	10
Data	1	0	0	1	,	A	1	,	H	R
REG_EXP	.	.	.	.	.	.	.	.	.	.

2.   & will hold 1001,A1,HR

Result of REG_EXP (.) = 1001,A1,HR

Replacement (&,) =1001,A1,HR,          

After the 3rd -e switch record will be: 1001,A1,HR,50000USA

4th -e switch will receive input as "1001,A1,HR,50000USA"

$ sed -e 's/./&,/10'

1.  Dot(.) will match until 16th occurrence of any character.

Occurrence	1	2	3	4	5	6	7	8	9	10	11	12	13	14	15	16
Data	1	0	0	1	,	A	1	,	H	R	,	5	0	0	0	0
REG_EXP	.	.	.	.	.	.	.	.	.	.	.	.	.	.	.	.

 2. & will hold 1001,A1,HR,50000USA

Result of REG_EXP (.) = 1001,A1,HR,50000

Replacement (&,) =1001,A1,HR,50000,

After the 4th -e switch record will be: 1001,A1,HR,50000,USA

Exactly what we want…!!!

Conclusion: We have already seen the same functionality using awk command, you may also check that out How to Convert Fixed Width File to Delimited Using AWK. & is a very powerful switch available in sed command to format the data. 

Keep Reading, Keep Learning, Keep Sharing...!!!