Port tests in timestamp.rs to sqllogictest. Part 1

[caicancai]

Which issue does this PR close?

part of #8216

Rationale for this change

What changes are included in this PR?

Are these changes tested?

Are there any user-facing changes?

[caicancai]

@alamb Hello,excuse me
I found that there is a test about the now function in timestamp.slt. I don’t know if I should add it.
And if I add the select now test, the result may be fixed, but the result of now is changing all the time, which may have an impact on the project itself.

[caicancai]

Dot't merge

[alamb]

@alamb Hello,excuse me
I found that there is a test about the now function in timestamp.slt. I don’t know if I should add it.
And if I add the select now test, the result may be fixed, but the result of now is changing all the time, which may have an impact on the project itself.

Hi @caicancai -- if there are tests that can't be ported to sqllogictest (for example, that now() doesn't change across session statements) can you please put a comment on them saying something like "can't write this in sqllogictest, so use a rust test" and just leave them in timestamp.rs?

Thank you 🙏

[caicancai]

@alamb Hello, I have two questions here.

1. The result I generated here seems to be different from the original result. Maybe there is something wrong with my table creation statement?
2. I have observed that there are many duplicate test samples in timestamp_coercion. I am not sure whether I should add them all. My idea is that I only need to add one test.

[comphead]

I think now can still be ported, its not stable value, but we can do it quasi stable like

select now() = current_time

or

select now() > to_date('2023-01-01')

[caicancai]

I think now can still be ported, its not stable value, but we can do it quasi stable like

select now() = current_time

or

select now() > to_date('2023-01-01')

Thank you for your reply. I noticed that the original timestamp.slt already has some tests belonging to now. I am not sure whether to continue adding them.

[comphead]

Thanks for double checking now() is in timestamp.slt, I also verified that and looks like test in slt for now is enough, and now I'm wondering if we still need timestamp.rs?

[alamb]

Thank you @caicancai -- this looks like a good improvement to me. Perhaps we can investigate @comphead 's question about needing timestamp.rs at all as part of a follow on PR

[caicancai]

Thanks for double checking now() is in timestamp.slt, I also verified that and looks like test in slt for now is enough, and now I'm wondering if we still need timestamp.rs?

From a personal point of view, I agree

[comphead]

lgtm @caicancai thanks.
I think it will be good to create a followup ticket to move remaining tests. I went trhough them real quick they looks migratable imho