usage of regex - sometimes I don't get the desired result, how to bring light on this black box?

[aisbergde]

SchemaCrawler is based on regex, but to get the right result is often a question of trial and error.

One finding is, that in some cases I have to use " (maybe also ' works) to get the expression into the parameter.

This doesn't work:

schemacrawler --url=jdbc:sqlserver://localhost;instancename=sql2019;databaseName=dhw_self; --schemas=dhw\_self\..* --tables=(?i)(dhw\_self\.repo\.generatorusp)|(dhw\_self\.repo\.generatoruspparameter)|(dhw\_self\.repo\.generatoruspstep) --info-level=standard --command=list --user=readonly --password=readonly

This works:

schemacrawler --url=jdbc:sqlserver://localhost;instancename=sql2019;databaseName=dhw_self; --schemas=dhw\_self\..* --tables="(?i)(dhw\_self\.repo\.generatorusp)|(dhw\_self\.repo\.generatoruspparameter)|(dhw\_self\.repo\.generatoruspstep)" --info-level=standard --command=list --user=readonly --password=readonly

Often the resulting set from the expression is bigger than the expected result

The problem concerns the shell in particular, but not only

some examples:

grep --grep-columns=dhw\_self\.repo\.RepoObject\..*
or
grep --grep-columns="dhw\_self\.repo\.RepoObject\..*"

In my naive imagination this should limit the object to dhw_self.repo.RepoObject, but the result is: I get all objects starting with 'dhw_self.repo.RepoObject', not only the one object. I get half the database because many objects start with "RepoObject".

#427

in the shell, when I use limit --schemas=WideWorldImporters\..* I get not only the schema WideWorldImporters, but I get also some object from WideWorldImportersDW. But it works when not in the shell.

schemacrawler.config.properties
When I try to filter out the shema sys using this:
schemacrawler.table.pattern.exclude=.*\.sys\..*
this will filter not only objects in the schema sys (I try to filter objects containing ".sys."), but also object containing sys anywhere, for example in the schema repo_sys
Idea was to exclude all system objects:
schemacrawler.table.pattern.exclude=.*\.sys\..*|.*\.INFORMATION_SCHEMA\..*
but I can't use this filter because it is filtering too much

black box

In other words, I try to find some regex and I test it in https://regex101.com/ or other places. And there I get the desired result. But in SchemaCrawler sometimes I get the expected result, but sometimes not. This is a very "black box" for me.

• It depends on: am I in the shell or not
• maybe the used scripting engine could also influence the result

Are there ways to shed some more light on this black box?
Could there be ways to see the actual regex expression applied?

[sualeh]

@aisbergde - please use SchemaCrawler 16.14.2