Shell Programming and Scripting

83510000000000063800000.1800000.1600000.1600000.2400000.1800000.2000000.21
83510000000000064500000.1100000.2800000.1000000.1200000.1900000.1200000.12
83510000000000064500000.1100000.2800000.1000000.1200000.1900000.1200000.13
83510000000000064900001.7800001.0400001.2300001.2600001.0900001.9000001.48
83510000000000065100000.2500000.1400000.2300000.3300000.2300000.2300000.20

sed -n "s/835100000000000638/835100000000000639/p oldfile > newfile

sed  -i.bak "s/835100000000000638/835100000000000639/g" urfile

perl -i.bak -pe 's/835100000000000638/835100000000000639/g' urfile

# strip lines out of file and make new file of lines removed and modified

# set up patterns - any regex will work

BEGIN {
        MY_PAT="835100000000000638"
        NEW_PAT="835100000000000639"
}

# This clause is executed on the opening of each file on the command line
# It first checks to see if this is not the first file and cleans up the
#  previous files and copies the new stripped file to the old file
#  NOTE:  I would prefer if it didn't overwrite the input file...
# It then creates the names and commands it will need later
(FNR == 1) {                    # check if first line of new file
        if ( NR != 1)           # see if this is not first file
        {
                close(FILEOUT)  # if true, then we close previous files
                close(FILELOG)
                system(FILECP)  # do the copy
                system(FILERM)  # remove the work file
        }

        FILEOUT = FILENAME ".strip"     # create striped file name using input file name
        FILELOG = FILENAME ".new"               # create new file

        FILECP = "cp " FILEOUT " " FILENAME     # copy command
        FILERM = "rm " FILEOUT          # remove strip file
}

# This clause executes for every line
# It copies lines from the input file to the appropriate output file
{
        if (  sub(MY_PAT,NEW_PAT) == 0 )      # test for pattern in line
                print > FILEOUT # save to stripped file if pattern not found
        else
                print > FILELOG # save to remove log if pattern match
}

# This clause is executed when the last line of the last file is reached
END {
        system(FILECP)  # clean up last file
        system(FILERM)
}

awk -f strip.awk <data.file>

$ cat a
83510000000000063800000.1800000.1600000.1600000.2400000.1800000.2000000.21
83510000000000064500000.1100000.2800000.1000000.1200000.1900000.1200000.12
83510000000000064500000.1100000.2800000.1000000.1200000.1900000.1200000.13
83510000000000064900001.7800001.0400001.2300001.2600001.0900001.9000001.48
83510000000000065100000.2500000.1400000.2300000.3300000.2300000.2300000.20

$ sed -i -e 's/835100000000000638/835100000000000639/gw test'  -e '/835100000000000639/d' a

$ cat test
83510000000000063900000.1800000.1600000.1600000.2400000.1800000.2000000.21

$ cat a
83510000000000064500000.1100000.2800000.1000000.1200000.1900000.1200000.12
83510000000000064500000.1100000.2800000.1000000.1200000.1900000.1200000.13
83510000000000064900001.7800001.0400001.2300001.2600001.0900001.9000001.48
83510000000000065100000.2500000.1400000.2300000.3300000.2300000.2300000.20

It is single sed command. 
$ sed -i -e 's/835100000000000638/835100000000000639/gw test'  -e '/835100000000000639/d'

sed: Not a recognized flag: i
Usage:  sed [-n] Script [File ...]
        sed [-n] [-e Script] ... [-f Script_file] ... [File ...]