Clarification in pipeline syntax

Help Connectors and Pipelines
1

I am referring below MemSQL url.

I need few clarification regarding below syntax.

[REPLACE | IGNORE | SKIP { ALL | CONSTRAINT | DUPLICATE KEY } ERRORS]

1)What is the difference between IGNORE & SKIP
2)What should I specify to handle errors,duplicate key violations and not null constraints
3)What is the difference between all vs errors
4)What is the difference between constraint and duplicate key errors?
5)My need is if any error occurs ,pipeline shouldn’t be stop and errors need to be captured in pipeline_erros tables.

2

Here is where we are going over the differences. IGNORE is MySQL compat and the recommended approach is SKIP