mjhjmtu

I have a big text file has lines with all numbers and with all letters and characters and also has lines mixed with numbers , letters and characters i want to print only the lines that have all numbers except the last or the last two characters.

1. 
   

   Print the line that start with number but ends with any character at the last or the last two characters that is not a digit.
   For example

   
   
   

   1234567a
   2245678902da
   A1234566d
   12345678abc
   

   
   
   

   The output have to be

   
   
   

   1234567a
   22345678902da
   

   
   


2. 
   

   Print the lines that have all their characters are digits except the first or the first and second characters to be not a digit.
   For example

   
   
   

   A1234
   Ab1234
   1a1234
   Abc1234
   

   
   
   

   The output have to be

   
   
   

   A1234
   Ab1234
   

   
   

Thanks

$ grep -Ex '[[:digit:]]+(.|[^[:digit:]]2)' file1
1234567a
2245678902da

The extended regular expression [[:digit:]]+(.|[^[:digit:]]2) will match one or more digits followed by either an unspecified character or two non-digits (this is a literal interpretation of your specification "start with number but ends with any character at the last or the last two characters that is not a digit"). The -x option to grep ensures that the match will be across full lines.

Note that this literal interpretation of what you specified also matches lines that contain only digits.

With

$ grep -Ex '[^[:digit:]]1,2[[:digit:]]+' file2
A1234
Ab1234

we match lines that start with one or two non-digits, and then contains one or more digits until the end of the line.

For a visual representation of the two regular expressions (and also at the same time showing how to do it with sed):

The first:

$ sed -nE 's/^([[:digit:]]+)(.|[^[:digit:]]2)$/(1)(2)/p' file1
(234567)(a)
(2245678902)(da)

The second:

$ sed -nE 's/^([^[:digit:]]1,2)([[:digit:]]+)$/(1)(2)/p' file2
(A)(1234)
(Ab)(1234)

Remove all parentheses from the sed command to get the sed solution.

Slightly different interpretation. I read this as the last 1-2 characters cannot be digits so I get

grep -E '^[[:digit:]]+[^[:digit:]]1,2$' file
1234567a
2245678902da

For first question below is command

sed -n '/^[0-9]*[a-z],2$/p' filename

output

1234567a
2245678902da

For second question below is the command

 sed -n '/^[A-Za-z]1,2[0-9]*$/p'

output:

A1234
Ab1234