-
Notifications
You must be signed in to change notification settings - Fork 69
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Add Result#columns
to store field types
#124
base: main
Are you sure you want to change the base?
Conversation
#undef XX | ||
}; | ||
|
||
static char *downcase(const char *str) |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
This is to convert the charset/type name into a lowercase symbol. I'm sure there is a better way to do this.
Trilogy_Result_Column, rb_intern("new"), 6, trilogy_result_columns_ctx->column_info[i].name, | ||
rb_id2sym(rb_intern(downcase(trilogy_type_names[trilogy_result_columns_ctx->column_info[i].type]))), | ||
rb_int_new(trilogy_result_columns_ctx->column_info[i].len), rb_int_new(trilogy_result_columns_ctx->column_info[i].flags), | ||
rb_id2sym(rb_intern(downcase(trilogy_charset_names[trilogy_result_columns_ctx->column_info[i].charset]))), |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Wrapping rb_id2sym(rb_intern(...))
seemed to be a way to get the actual string-like symbol value i.e. :long
. There is probably a better way to do this but I couldn't find one 🤷
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Does this leak memory? Afaik rb_intern
doesn't take ownership of the allocated memory but ends up copying it?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Thanks - given the fact that this already felt sub-optimal, I wonder if it makes more sense to just set an int here and then cast to the symbol on the Ruby class.
contrib/ruby/lib/trilogy/result.rb
Outdated
|
||
|
||
class Columns | ||
include Enumerable |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Do we need this to be Enumerable
? It doesn't look like we return a Columns
object anywhere (it's an intermediate so we have a place to hold the raw column data)—rather we call all
on it and return an Array
.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Good call - you are correct that we don't need this. Removed and updated the previous commit
(As an aside, not sure what the preference is for commit history here. Happy to squash these three down to a single commit if we prefer)
401a44f
to
d27a2c6
Compare
Towards trilogy-libraries#99, this exposes additional field data on the query result via a `Result#columns` method. We were already using this data in the C extension to determine how to cast the column values. We now add a pointer to the `column_info` struct onto the result and then lazily add the values to the `Result` instance when called. This method could eventually replace the `#fields` method which at the moment only includes the column name. Co-authored-by: Daniel Colson <[email protected]> Co-authored-by: Charlotte Wen <[email protected]>
We want to set some of these values as symbols instead of ints on the Column to make it easier to work with. In this commit, we created loopups for int to string using macros, then changed the type to a symbol via `rb_intern` + `rb_id2sym`. It's possible there's a single function to do this, but we couldn't find one. I also added a downcase function to make sure we get the values in lower instead of uppercase. I'm sure there is a better way to do this. I have no idea how to write C. Co-authored-by: Daniel Colson <[email protected]>
Previously, we had added a pointer to our column_info struct onto our `Result` for lazy loading the data. This resulted in changing the type of `Result` to `T_DATA` which could have negative performance implications. Instead, we can create a separate `Columns` class to store the pointer to our struct. A reference to `columns` now gets eagerly loaded onto the result, but we only populate the actual information for the columns when calling `Result#columns`. This seems like a good trade off for lazy loading (vs. putting the pointer directly on `Result`). Having said that, it's probably worth asking if the lazy loading is worth this extra complexity at all, or if we shoudl simply populate `@columns` with an already loaded array for the `Result` immediately.
d27a2c6
to
1131f77
Compare
struct trilogy_result_columns_ctx { | ||
struct column_info *column_info; | ||
uint64_t column_count; | ||
}; |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
struct trilogy_result_columns_ctx { | |
struct column_info *column_info; | |
uint64_t column_count; | |
}; | |
struct trilogy_result_columns_ctx { | |
uint64_t column_count; | |
struct column_info column_info[]; | |
}; |
You could use a flexible array member to allocate the whole list at once and avoid pointer chasing and malloc churn.
Closes #99
This PR adds a new
Result#columns
method which returns an array ofResult::Column
objects to represent column information such as name, length, and type. At some point this could replaced thefields
method, but as of now was added as a non-breaking change.We added a new
Columns
class as well. When instantiating a newResult
in the C extension, we assign a new instance of ourColumns
class which carries a pointer to a struct. After populating the properties of the result, we cache thecolumn_info
in a struct on theColumns
Value and add the information in (including casting some integers into symbol values). We can then lazily-load the data out of the struct and populate the ruby array when calling#columns
Full disclosure: I've never written a C extension (or any C for that matter) before. Any and all feedback welcome - I added a few inline comments in the PR.
CC @composerinteralia @ywenc