Conversation
|
The main purpose is to be able to use a transaction like [{:db/ident :my.attr
:db/valueType :db.type/ref}] and have it recognized in datahike as part of the schema (which stores schema as a plain map). It also enables all sorts of middleware for transactions which can allow for customization based on different performance needs. I've decided against using transducers to replace This patch is a little more fully formed than tonsky/datascript#234 with some tests. The efficiency could be sped up if schema changes did not require re-creating the indexes. The next step there would be completing the schema validation function. |
No description provided.