DuckDB 1.2.0

Author: AlexR2D2

CHANGELOG.md

@@ -1,5 +1,10 @@
 # Changelog
 
+0.3.9
+- [DuckDB 1.2.0 release](https://github.com/duckdb/duckdb/releases/tag/v1.2.0). Please, read the [Announcing DuckDB 1.2.0](https://duckdb.org/2025/02/05/announcing-duckdb-120)
+- Fixed the isinf/isnan build error on Linux
+- `DuckDB.query(sql)` without parameters can execute multiple SQL statements at onсe.
+
 0.3.8
 - Added transaction managing functions: begin_transaction, commit, rollback, set_auto_commit, is_auto_commit, has_active_transaction
 

README.md

@@ -12,9 +12,7 @@ The library uses `amalgamation` of the DuckDB sources, which combine all sources
 
 All NIF functions implemented as [**Dirty NIF**](https://www.erlang.org/doc/man/erl_nif.html)
 
-**DuckDB is in the active development phase, and new version of library may not open the database created by the old version.**
-
-[HexDocs](https://hexdocs.pm/duckdbex/)
+Online [HexDocs](https://hexdocs.pm/duckdbex/)
 
 ## Installation
 
@@ -24,7 +22,7 @@ by adding `duckdbex` to your list of dependencies in `mix.exs`:
 ```elixir
 def deps do
   [
-    {:duckdbex, "~> 0.3.0"}
+    {:duckdbex, "~> 0.3.9"}
   ]
 end
 ```
@@ -401,3 +399,27 @@ conf = %Duckdbex.Config{allow_unsigned_extensions: true}
 ```
 
 Documentation generated with [ExDoc](https://github.com/elixir-lang/ex_doc)
+
+## Huge numbers (hugeint)
+
+The BIGINT and HUGEINT types are designed to be used when the range of the integer type is insufficient. Hugeint in DuckDB is 128bit integer. Because max native integer in C++ is 64bit integer, HUGEINT is represented as combination of two 64bit integers. There is no efficient way to pass HUGEINT from DuckDB(C++) to Elixir(Erlang) than to pass it as is - combination of two 64bit integers. But in Elixir there is no restrictions to integers, so if you will get instead of integer a tuple of two integers you should conver it to integet via `DuckDB.hugeint_to_integer({upper_int, lower_int})`.
+
+```elixir
+> {:ok, r} = Duckdbex.query(conn, "SELECT SUM(1);")
+> Duckdbex.fetch_all(r)
+[[{0, 1}]]
+> Duckdbex.hugeint_to_integer({0, 1})
+1
+```
+
+And vice versa, if you should pass HUGEINT as argument to sql query, you should convert this argument to 'native' representation
+
+```elixir
+> hi = Duckdbex.integer_to_hugeint(123456789123456789123456789)
+{6692605, 17502027875430457109}
+> {:ok, r} = Duckdbex.query(conn, "SELECT SUM(1234567891234567891234567891) > $1;", [hi])
+> Duckdbex.fetch_all(r)
+[[true]]
+```
+
+Currently Duckdbex lib didn't convert automatically `hugeint_to_integer` for you because this is additional extra pass through your collection of rows which will be executed inside the library.