Delete lines in a file that start with specific strings - sed

Question

I have some files that look like this:
Node Present
1 243
2 445
10 65
4 456
43 8
...
I need to remove the values corresponding to specific nodes and I have a file specifying this nodes that looks like this:
1
4
...
The idea is to delete the lines that start with the values specified in my second file. I know that "sed" can do something like this, but I do not know how to apply it for all the values specified in the second file. More over, I want to delete node 1, but not node 100, and I am seeing that node 100 will also get erased with my approach.
sed '/^1/d'

Answer 1

This sort of problem is typically done with awk, and is quite common. Read the first file into an array, and then use it to process subsequent files. For example:
$ cat skip
1
4
$ cat input
Node Present
1 243
2 445
10 65
4 456
43 8
...
$ cat input2
Node Present
1 243
2 445
4 456
43 8
4 587
...
$ awk 'NR==FNR{a[$1] = 1; next} ! a[$1] && FNR > 1' skip input input2
2 445
10 65
43 8
...
2 445
43 8
...
The initial NR == FNR causes those command to only be executed on the first file, loading the array with the ids you wish to skip. Subsequent commands print lines in which the first column did not appear in the first file and the first line of each file, in which FNR > 1. (FNR is the "File Record Number", aka the line number in the file.)

Answer 2

sed is not the right tool for this job. I suggest using awk like this:
awk 'NR == FNR {ids[$1]; next} NR == 1 || !($1 in ids)' ids nodes
Node Present
2 445
10 65
43 8
Where input files are:
cat ids
1
4
cat nodes
Node Present
1 243
2 445
10 65
4 456
43 8

Answer 3

If you want to stick with sed and bash, here is one possible way how you would do it. Please note that, for large files, that would take orders of magnitude longer to run than the awk answers (similar to time complexity of linear search vs binary search).
sed -f <(sed 's/^/\/^/;s/[[:blank:]]*$/[[:blank:]]\/d/' ids) nodes
ids is filename of single-column file and nodes is that of two-column.