39 lines
1.1 KiB
Bash
39 lines
1.1 KiB
Bash
|
#!/bin/bash
|
||
|
|
|
||
|
|
if [ $# -eq 2 ]
|
||
|
|
then
|
||
|
|
OLD=$1
|
||
|
|
NEW=$2
|
||
|
|
|
||
|
|
# We need a tab character as a field separator
|
||
|
|
TAB=`echo -e "\t"`
|
||
|
|
|
||
|
|
#Temporary storage
|
||
|
|
TEMPFILE=`mktemp /tmp/catmerge.XXXXX`
|
||
|
|
|
||
|
|
# Extract the list of keys to remove
|
||
|
|
# Compare (diff) the keys only (cut) ; keep only 'removed' lines (grep -),
|
||
|
|
# Ignore diff header and headlines from both files (tail), remove diff's
|
||
|
|
# prepended stuff (cut)
|
||
|
|
# Put the result in our tempfile.
|
||
|
|
diff -u <(cut -f 1,2 $OLD) <(cut -f 1,2 $NEW) |grep ^-|\
|
||
|
|
tail -n +3|cut -b2- > $TEMPFILE
|
||
|
|
|
||
|
|
# Reuse the headline from the new file (including fingerprint). This gets
|
||
|
|
# the language wrong, but it isn't actually used anywhere
|
||
|
|
head -1 $NEW
|
||
|
|
# Sort-merge old and new, inserting lines from NEW into OLD (sort);
|
||
|
|
# Exclude the headline from that (tail -n +2)
|
||
|
|
# Then, filter out the removed strings (fgrep)
|
||
|
|
sort -u -t"$TAB" -k 1,2 <(tail -n +2 $OLD) <(tail -n +2 $NEW)|\
|
||
|
|
fgrep -v -f $TEMPFILE
|
||
|
|
|
||
|
|
rm $TEMPFILE
|
||
|
|
|
||
|
|
else
|
||
|
|
echo "$0 OLD NEW"
|
||
|
|
echo "merges OLD and NEW catalogs, such that all the keys in NEW that are"
|
||
|
|
echo "not yet in OLD are added to it, and the one in OLD but not in NEW are"
|
||
|
|
echo "removed. The fingerprint is also updated."
|
||
|
|
fi
|