orm: null handling and optional fields #19379
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Nullable DB fields are now determined by corresponding option struct field. The [nonull] attribute is deprecated and fields are all NOT NULL now, unless they are option fields. New orm primitive, NullType, added to support passing none values to db backends, which have been updated to support it. Also, empty string and 0 numberic values are no longer skipped during insert (since they may be valid values).
also, improved formatting for orm cgen, and removed optimised operand handling of orm `is` and `!is` operators
During insert, fields which are * [sql: serial] * [default: whatever] and where the data is a default value (e.g., 0, ""), those fields are not sent to the db, so that the db can generate auto-increment or default values. (This was previously done only for [primary] fields, and not in all circumstances, but that is not correct -- primary and serial/auto-increment fields are differnet.)
Changed orm.write_orm_select() so that you pass to it the name of a resut variable which it populates with the result (or not) and changed use of it in sql_select_expr() and calls in write_orm_select() to populate substructs.
|
I am working through resolving conflicts with changes made to master... 😢 |
|
This is great stuff! Thanks, Tim. |
JalonSolov
reviewed
Sep 18, 2023
edam
force-pushed
the
pr-orm-null-handling
branch
from
311bb0b
to
3532ab7
Compare
Code in checker/orm.v used the SqlStmtLine object field name to store c-specific referenecs to option and array fields (for arrays of children). I moved this logic to cgen. And fixed an issue introduced with option fields, where an array of children was unpacked into a non-array result which could corrupt memory.
edam
force-pushed
the
pr-orm-null-handling
branch
from
50d5e43
to
c1df21a
Compare
edam
changed the title
spytheman
force-pushed
the
pr-orm-null-handling
branch
from
a306df7
to
8b027b4
Compare
…tes.vv; run `VAUTOFIX=1 ./v vlib/v/compiler_errors_test.v`
spytheman
approved these changes
Oct 5, 2023
Wertzui123
pushed a commit
to Wertzui123/v
that referenced
this pull request
Oct 8, 2023
* orm: added is none and !is none handling * orm: added NullType, support option fields and deprecate [nonull] Nullable DB fields are now determined by corresponding option struct field. The [nonull] attribute is deprecated and fields are all NOT NULL now, unless they are option fields. New orm primitive, NullType, added to support passing none values to db backends, which have been updated to support it. Also, empty string and 0 numberic values are no longer skipped during insert (since they may be valid values). * orm: fix [nonull] deprecation warning * orm: add null handling to update and select also, improved formatting for orm cgen, and removed optimised operand handling of orm `is` and `!is` operators * sqlite: read/report NULLs using new orm NullType * postgres: returning data primitives now returns new orm.NullType * orm: initialise NullType Primitives properly * orm: do not smart cast operands inside sql * orm: fix bad setting of option value * orm: improve orm_null_test.v, adding/fixing selects * orm: cleanup: rename NullType->Null, use serial const, cgen output * orm: handle automatically generated fields more explicitly During insert, fields which are * [sql: serial] * [default: whatever] and where the data is a default value (e.g., 0, ""), those fields are not sent to the db, so that the db can generate auto-increment or default values. (This was previously done only for [primary] fields, and not in all circumstances, but that is not correct -- primary and serial/auto-increment fields are differnet.) * orm: udpated README * orm: select cgen fixes: read from uninit res; fail to init res * orm: udpated tests * orm: fix option sub-struct fields * orm: fixed joins to option structs Changed orm.write_orm_select() so that you pass to it the name of a resut variable which it populates with the result (or not) and changed use of it in sql_select_expr() and calls in write_orm_select() to populate substructs. * orm: fix pg driver handling of NULL results * orm: move runtime checks to comptime checker; cache checked tables * orm: vfmt :( * orm: markdown formatting * orm: renamed orm.time_ and orm.enum_; updated db drivers * checker: updated orm tests * orm: fix issue setting up ast option values as orm primitives * checker: ORM use of none/options and operations (added tests) * orm: fixed tests * db: clean code * examples: remove orm nonull attributes * orm: skip test memory santisation for orm_null_test.v * orm: make the type-to-primitive converstion fns not public * orm: mv object var c-code from checker->cgen; fix memory corruption Code in checker/orm.v used the SqlStmtLine object field name to store c-specific referenecs to option and array fields (for arrays of children). I moved this logic to cgen. And fixed an issue introduced with option fields, where an array of children was unpacked into a non-array result which could corrupt memory. * orm: fixed vast error * orm: skip 2 tests on ubuntu-musl which require sqlite3.h * cgen: prevent casting a struct (string) * v fmt orm_fkey_attribute.vv, orm_multidim_array.vv, orm_table_attributes.vv; run `VAUTOFIX=1 ./v vlib/v/compiler_errors_test.v`
juan-db
added a commit
to juan-db/v
that referenced
this pull request
Dec 28, 2023
The @[nonull] attribute is no longer used and is replaced by using an option type. vlang#19379
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This PR adds:
NOT NULLand unless they are option fieldsorm.Nulltypeis noneand!is nonein WHERE clausesnone(e.g.,update Foo, set a = none where...)serialfields and fields withdefaultattr are not INSERTedApologies for lumping so much stuff together! They are related changes. (I.e., I couldn't implement NULL handling without option fields.)
🤖 Generated by Copilot at 377b1a3
This pull request enhances the ORM module by generating more readable and consistent C code, and by supporting more features and types for database operations. It also refactors some of the ORM code generation logic for clarity and simplicity.
🤖 Generated by Copilot at 377b1a3
isandnot_isoperations, and handlingoptiontypes andtime.Timetype in the data array and the result variable (link, link, link, link)find_first,contains, andcontains_argmethods instead of looping over the attributes manually (link, link)