Add getting column specification by name #276
Conversation
psarna
force-pushed
the
add_getting_spec_by_name
branch
from
2dc1cc7 to
59d90e9
Compare
|
(some extra unrelated commits were initially pushed by mistake, fixed) |
psarna
force-pushed
the
add_getting_spec_by_name
branch
from
59d90e9 to
039b98b
Compare
examples/get_by_name.rs
Outdated
| .query("SELECT pk, ck, value FROM ks.hello", &[]) | ||
| .await?; | ||
| let (ck_idx, _) = query_result | ||
| .get_spec_by_name("ck") |
There was a problem hiding this comment.
The get_spec_by_name function name looks confusing to my 👀. Perhaps something like get_column_spec is more readable?
There was a problem hiding this comment.
Sure. I added _by_name in case we ever want to get the spec by index, but for that we can just use the vector directly. I'll change to get_column_spec
| @@ -75,3 +75,7 @@ path = "clone.rs" | |||
| [[example]] | |||
| name = "speculative-execution" | |||
| path = "speculative-execution.rs" | |||
|
|
|||
There was a problem hiding this comment.
(I can't attach this comment to the commit message, so I'm placing it here)
"It's a first step towards a macro which gets the value by name straight from a result row." - is the commit message correct? If the answer is yes then could you elaborate on the macro idea?
There was a problem hiding this comment.
If we also decide to add column specification information to every row (which is a waste, but to consider), then it would be even more convenient to extract values by name. It will have a performance penalty though, so I'm not 100% sure if we want to proceed with that direction
There was a problem hiding this comment.
OK, I see. I don't think a macro would be required in that case, we could just add a method (e.g. get_by_name) or even overload the indexing operator to work on strings for maximum ergonomics (see the std::ops::Index trait).
I'm not sure about adding column specification per row either. We don't have to clone it for each row, it could keep it behind an atomic pointer (Arc) so the cost is not as tragic as an e.g. allocation.
You're right, column spec should probably become part of
|
psarna
force-pushed
the
add_getting_spec_by_name
branch
from
039b98b to
94181ee
Compare
|
v2:
|
This information will be needed to implement getting column values by column name instead of using indexes.
With a column spec and its ordinal number, it will be possible to get specific columns out of the result.
As opposed to getting columns just by their indexes, it's now slightly easier to get a column value by its name. It's a first step towards a macro which gets the value by name straight from a result row.
One of the test cases is slightly amended to also check the mechanism which retrieves column specification from the query result.
psarna
force-pushed
the
add_getting_spec_by_name
branch
from
94181ee to
00fabbf
Compare
This series is a first baby step towards solving #271 - which is about enabling users to get column values from result rows by name instead of by index.
Pre-review checklist
Fixes:annotations to PR description.