Parsing csv and conditions

[Khuman]

I'm using elastic stack 7.3.2. I need to use working logstash config on new version. This one construction work perfect on version 6.5.0, but don't work on 7.3.2
filter {
csv {
columns => ["column1","column2","column3","column4"]
separator => ", "
quote_char=> "'"
source => "log"
}
if [column2] == "" {
drop{}
}
if ![column3] {
drop{}
}
}
But in my new index still appear documents with empty column3 and column2.
Can someone correct me?

[a.kuz (Anton)]

Hi. Probably i think you could reach this by:

Summary

filter {
	csv {
			columns => ["column1","column2","column3","column4"]
			separator => ", "
			quote_char=> "'"
			source => "log"
		}
		if [column2] ~= "" and [column3] {
			mutate {
				add_tag => [ "true" ]
			}
		}
	}
}

output {
	if [tags] == "true" {
        elasticsearch {
            ...
        }
	}
}