Antoine Fabri
Antoine Fabri
If I were to build a dm from a folder of csvs I think I would like an option to be guided through the process, maybe for each table we...
Maybe we could still produce (and try to read) a metadata file or sheet, if it's not there when reading we fall back on (optionally guided) automation and can even...
A few comments @krlmlr * Some DBMS limit names to as few as 30 bytes (e.g. Oracle 12.1 and below), your proposal makes this limit very easy to reach *...
All of the above doesn't mention primary keys. To reconstruct robustly without ptype we need to tag them too. I propose to suffix col names of primary keys with `*`.
Unnamed arguments in dm already work ``` r dm ── Metadata ──────────────────────────────────────────────────────────────────── #> Tables: `iris` #> Columns: 5 #> Primary keys: 0 #> Foreign keys: 0 ``` This doesn't however...
See also : https://github.com/cynkra/dm/issues/859 I also feel that `dm_add_tbl(dm, b = head(a))` should work. But I'd like to understand why we need `dm_add_tbl()` at all rather than generalising `dm_mutate_tbl()` so...
FWIW I took a very similar approach for {flow}'s hex logo :  The database icon might be used too, we could imagine something like below but with the database...
Since on remote databases we might be forced to include metadata in column names, maybe we could just suffix those with `_nested` or `_packed` ? Before wrapping we would check...
Is it better than using `DiagrammeRsvg::export_svg()` with `test that::expect_snapshot_file()` ? (or expect_snapshot()` on the xml content of the svg file ?)
> Is there an inverse of nest_join() ? No, we have only dm_unnest_tbl(). We can implement `unnest_unjoin()`, it might return a list of 2 tables : * "parent" : select...