crud
crud copied to clipboard
Easy assess to data stored in vshard cluster
`*_many` operation provides `operation_data` with each error so it would be possible to - track which exact record had been failed and - retry failed records afterward. But `operation_data` that...
```bash ./doc/playground.lua ``` ```lua tarantool> crud.min('customers', nil, {fields = {'unknown'}}) --- - error: '...eorgy/Development/github/tarantool/crud/crud/borders.lua:171: attempt to index local ''result'' (a nil value)' ... ```
**Added a way to search by multiple values:** ### Sample: crud.select('users', {{'=', 'name', **{'Andrey', 'Sergey'}**}}, {fields = {'_id', 'name'}}); ### Added new **operators** such as: '' - not '>>
`crud` uses `tuple-merger` for select. Built-in merger is provided for the following versions https://github.com/tarantool/crud/blob/2d3d47937fd02d938424659bc659fdc24a32dc8a/crud/common/utils.lua#L589-L599 and external merger can be provided for the following versions https://github.com/tarantool/crud/blob/2d3d47937fd02d938424659bc659fdc24a32dc8a/crud/common/utils.lua#L607-L620 There is a separate select...
This patch makes `tuple-merger` and `tuple-keydef` dependency non-optional. See commits for more info. Closes #365
We need to support tests with a master switch with `vshard`: 1. `test/integration/select_readview_test.lua.test_select_switch_master` 2. `test/integration/select_readview_test.lua.test_select_switch_master_first` But there are [no proper switchover](https://github.com/tarantool/crud/pull/381#discussion_r1347504713) for `vshard` so we have to figure out how...
After #364, perf tests length is ~1h10m. One must wait till perf tests succeed each time to get an `ok` from test CI workflow. On the other hand, we don't...
Follows up #372 crud statistics may cover the count of `crud.readview` calls, as well as the count of current open readviews.
`crud.pairs` has `use_tomap` option allowing to return tuples as key-value maps instead of arrays, since there is no way to do `unflatten_rows` while calling pairs. But this feature seems useful...
There is a problem with cursor based pagination on requests by secondary indexes. Consider a simple key-value schema: | **field** | **type** | **info** | |------------|------------|-----------| | **key** | string...