NetNews Usenet Archive 1992 #30/NN_1992_30.iso/spool/comp/sys/next/misc/23024/cut

Text File | 1992-12-14 | 4.0 KB | 182 lines

#!/bin/sh
# Shell/awk version of System V cut(1) command
# Is not fully compatible, nor optimized (-c option is SLOW).
# Incompatibilities include
#  - Different diagnostics
#  - Arbitrary limits on number of columns/fields it will output
#  - Does not process delimiter-less records properly
#  - Does not implement the -s option

# Written by Jon LaBadie (princeton!jonlab!jon) in July, 1990 while
# teaching a UNIX Tools class on a Xenix system that did not include
# cut.  The command was needed for some of the class exercises.

# In no way would I recommend it be used as a replacement for cut(1).
# But if you need the capabilities of cut, this might do.

# set -u

PROG=${0}

usage() {
    # usage message should be revised to include -d option
    echo "Usage:  ${PROG} -{f|c}spec [file(s)]" >&2
    echo "  where:  spec is m,n or m-n, or n-" >&2
    exit 1
}

COLUMNS=FALSE
FIELDS=FALSE
DELIM=FALSE

while [ ${#} -gt 0 ]
do
    case "${1}"
    in
    -f*)    [ ${COLUMNS} != FALSE ] && usage
        FIELDS=`expr "${1}" : '..\(.*\)'`
        [ "${FIELDS}" = "" ] && {
            echo "${PROG}: -f option requires an argument" >&2
            usage
        }
        ;;
    -c*)    [ ${FIELDS} != FALSE ] && usage
        COLUMNS=`expr "${1}" : '..\(.*\)'`
        [ "${COLUMNS}" = "" ] && {
            echo "${PROG}: -c option requires an argument" >&2
            usage
        }
        ;;
    -d)    echo "${PROG}: -d option requires a single character argument" >&2
        usage
        ;;
    -d?)    DELIM=`expr "${1}" : '..\(.*\)'`
        ;;
    -d??*)    echo "${PROG}: single character delimiters only" >&2
        exit 1
        ;;
    -?*)    echo "${PROG}: Illegal Option: \"${1}\"" >&2
        usage
        ;;
    *)    break    # start of file names
        ;;
    esac
    shift
done

[ "${FIELDS}" = FALSE -a "${COLUMNS}" = FALSE ] && {
    echo "${PROG}: either -f or -c option must be used" >&2
    usage
}

[ "${DELIM}" != FALSE -a "${COLUMNS}" != FALSE ] && {
    echo "${PROG}: -d and -c options are incompatible" >&2
    usage
}

[ "${DELIM}" = FALSE ] && DELIM='    '   # TAB character

FILES=${@+"${@}"}

if [ "${FIELDS}" != FALSE ]
then
    awk -F"${DELIM}" '
    BEGIN {
        OFS = "'"${DELIM}"'"
        stderr = "/dev/tty"

        n = split("'"${FIELDS}"'", tmp1, ",")
        j = 0

        for (i = 1; i <= n; i++) {
            if ( (pos = index(tmp1[i], "-")) == 0) {
                # must be simple n type column spec
                cols[++j] = tmp1[i] + 0;
            }
            else if (pos == length(tmp1[i])) {
                # must be n- type column specifier
                if (i != n) {
                    print "'"${PROG}"'" ": Field specifier " tmp1[i] ", must be last" > stderr
                    exit 1
                }
                v = substr(tmp1[i], 1, pos - 1) + 0
                while (j < 48)  #  arbitrary max of 48
                    cols[++j] = v++
            }
            else {
                # last possibility is m-n type
                n2 = split(tmp1[i], tmp2, "-")
                while (tmp2[1] <= tmp2[2])
                    cols[++j] = tmp2[1]++ + 0
            }
        }
        ncols = j
        for (i = 1; i < ncols; i++)
            if (cols[i] >= cols[i+1]) {
                print "'"${PROG}"'" ": Fields must be specified in ascending order" >stderr
                exit 1
            }
    }

    {
        printf("%s", $(cols[1]))
        for (i = 2; i <= ncols && cols[i] <= NF; i++) {
            printf("%s%s", OFS, $(cols[i]))
        }
        print ""
    }
    ' ${FILES}
else
# must be -c option
    awk '
    BEGIN {
        stderr = "/dev/tty"

        n = split("'"${COLUMNS}"'", tmp1, ",")
        j = 0

        for (i = 1; i <= n; i++) {
            if ( (pos = index(tmp1[i], "-")) == 0) {
                # must be simple n type column spec
                cols[++j] = tmp1[i] + 0;
            }
            else if (pos == length(tmp1[i])) {
                # must be n- type column specifier
                if (i != n) {
                    print "'"${PROG}"'" ": Column specifier " tmp1[i] ", must be last" > stderr
                    exit 1
                }
                v = substr(tmp1[i], 1, pos - 1) + 0
                while (j < 200)  #  arbitrary max of 200
                    cols[++j] = v++
            }
            else {
                # last possibility is m-n type
                n2 = split(tmp1[i], tmp2, "-")
                while (tmp2[1] <= tmp2[2])
                    cols[++j] = tmp2[1]++ + 0
            }
        }
        ncols = j
        for (i = 1; i < ncols; i++)
            if (cols[i] >= cols[i+1]) {
                print "'"${PROG}"'" ": Columns must be specified in ascending order" >stderr
                exit 1
            }
    }

    {
        printf("%s", substr($0, cols[1], 1))
        l = length
        for (i = 2; i <= ncols && cols[i] <= l; i++) {
            printf("%s", substr($0, cols[i], 1))
        }
        print ""
    }
    ' ${FILES}
fi

exit 0