Kirill Müller

- [ ] Avoid proactive checks if a schema exists, let the database determine that check and fail accordingly - [ ] Use `dm_meta()` from #517 for `db_schema_list()` and `db_schema_exists()`...

Define data model for INFORMATION_SCHEMA and expose it

2

The `INFORMATION_SCHEMA` is a standardized way for databases to provide information on tables, columns and keys. See https://dataedo.com/kb/databases/all/information_schema for details. To better understand it, we should build a dm for...

Review real use cases for `dm_disentangle()`

5

in the order of complexity: - [ ] flights with cycle - [ ] financial with cycle (#1063) - [ ] customer database - [ ] ... - [ ]...

We want to split the process into "planning" and "doing", just like `dm_wrap_tbl()`, to be able to guess a plan if only the result of `dm_disentangle()` is given (e.g. `got_chars`...

Like with `dm_nycflights13()`. https://relational.fit.cvut.cz/dataset/Financial

- [ ] Disentangle - [ ] Specification - [ ] Inverse

for full automation. Example: `dm_unwrap_tbl(got_chars, ...)` once we can auto-unwrap.

Inverse of `dm_disentangle()`

4

if the user provides ptype.

Export `dm_disentangle()`

1

and document it. Review interface. Think about recipe and ptype.

``` r library(conflicted) library(dm) schemata % dm::dm_add_fk(tables, c(table_catalog, table_schema), schemata) %>% dm::dm_add_fk(columns, c(table_catalog, table_schema, table_name), tables) %>% dm::dm_add_fk(table_constraints, c(table_catalog, table_schema, table_name), tables) %>% dm::dm_add_fk(key_column_usage, c(constraint_catalog, constraint_schema, constraint_name), table_constraints) %>% dm::dm_add_fk(constraint_column_usage,...