deltalake-rb 0.1.2 → 0.1.3

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: cd4f4c8d4d173c5b44a1758ba798624c43724e702941f3044763d762ee53681e
-  data.tar.gz: 0c1616f7555ce13dc5d24837b6c8b47148e9a5308e615dee69d9b8b5f58b2116
+  metadata.gz: 74cdc17a6f4237378f874c7c2142ef5bcd233d7da9273296c5e2dbb24d925add
+  data.tar.gz: ddf786922725bddc6adeadeaa9e734d3c3cd974ba009e319740393ade76a44a6
 SHA512:
-  metadata.gz: f6fdce2f189671bb76328f7c3ef4cc2405897bafae9b3f31cfcbe95a60842890d643ad8e4da7af773a3e10ea993b0c7be86e3de5ff5ea836d523872e670fb25e
-  data.tar.gz: 728f32cf6c891ee82cb2690248197841c8a6b561d6fc4319969799117afafc1d6ea355782a7d906c07d65ca525da28c96a22051e0213dd8d09cbbaa410b07ed0
+  metadata.gz: d97802a6b1443cd1b85802fae78e81eff9657fe2b524631cc5a3cb4a09d18f01c408af388f52bad37a23b78dbc4816e605ae3690ef013ec904b6e47fd0a88ee0
+  data.tar.gz: fe3ffa3a6babd7fd154670ab587c77febb3be4f801d2a5421be373b6c165a198904f50980df7ef79e7dab155b01dddac6cba52a7f1d07225087462812c2996f0

data/CHANGELOG.md

@@ -1,3 +1,9 @@
+## 0.1.3 (2024-12-28)
+
+- Updated `deltalake` to 0.22.3
+- Added support for Ruby 3.4
+- Added `rechunk` and `columns` options to `to_polars` method
+
 ## 0.1.2 (2024-12-03)
 
 - Updated `deltalake` to 0.22.2

data/Cargo.lock

@@ -1572,13 +1572,13 @@ dependencies = [
 
 [[package]]
 name = "deltalake"
-version = "0.1.2"
+version = "0.1.3"
 dependencies = [
  "arrow",
  "arrow-schema",
  "chrono",
  "delta_kernel",
- "deltalake 0.22.2",
+ "deltalake 0.22.3",
  "futures",
  "magnus",
  "num_cpus",
@@ -1589,9 +1589,9 @@ dependencies = [
 
 [[package]]
 name = "deltalake"
-version = "0.22.2"
+version = "0.22.3"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "950788ee777d7fa22043fd329854809cbbf418d25dc04456509e56e885973a7a"
+checksum = "9e289d4760a09c95d44b978363b4466b0aacf7c82561e773f2f12eb5bbf257e9"
 dependencies = [
  "deltalake-aws",
  "deltalake-azure",
@@ -3174,18 +3174,18 @@ dependencies = [
 
 [[package]]
 name = "rb-sys"
-version = "0.9.103"
+version = "0.9.105"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "91dbe37ab6ac2fba187480fb6544b92445e41e5c6f553bf0c33743f3c450a1df"
+checksum = "4b3a1f3ce8e7c36d777d52fe7a99039fe4fea7c8ec355a4c4f3a17f92a14029f"
 dependencies = [
  "rb-sys-build",
 ]
 
 [[package]]
 name = "rb-sys-build"
-version = "0.9.103"
+version = "0.9.105"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "c4d56a49dcb646b70b758789c0d16c055a386a4f2a3346333abb69850fa860ce"
+checksum = "3e6b246c29c0809e1cbe60a1ba9e093da72a4676d02adc68469297d1e589bbf0"
 dependencies = [
  "bindgen",
  "lazy_static",

data/README.md

@@ -21,15 +21,15 @@ It can take 5-10 minutes to compile the gem.
 Write data
 
 ```ruby
-df = Polars::DataFrame.new({"a" => [1, 2], "b" => [3.0, 4.0]})
-DeltaLake.write("./data/delta", df)
+df = Polars::DataFrame.new({"id" => [1, 2], "value" => [3.0, 4.0]})
+DeltaLake.write("./events", df)
 ```
 
 Load a table
 
 ```ruby
-dt = DeltaLake::Table.new("./data/delta")
-df2 = dt.to_polars
+dt = DeltaLake::Table.new("./events")
+df = dt.to_polars
 ```
 
 Get a lazy frame
@@ -41,31 +41,31 @@ lf = dt.to_polars(eager: false)
 Append rows
 
 ```ruby
-DeltaLake.write("./data/delta", df, mode: "append")
+DeltaLake.write("./events", df, mode: "append")
 ```
 
 Overwrite a table
 
 ```ruby
-DeltaLake.write("./data/delta", df, mode: "overwrite")
+DeltaLake.write("./events", df, mode: "overwrite")
 ```
 
 Add a constraint
 
 ```ruby
-dt.alter.add_constraint({"a_gt_0" => "a > 0"})
+dt.alter.add_constraint({"id_gt_0" => "id > 0"})
 ```
 
 Drop a constraint
 
 ```ruby
-dt.alter.drop_constraint("a_gt_0")
+dt.alter.drop_constraint("id_gt_0")
 ```
 
 Delete rows
 
 ```ruby
-dt.delete("a > 1")
+dt.delete("id > 1")
 ```
 
 Vacuum
@@ -83,13 +83,13 @@ dt.optimize.compact
 Colocate similar data in the same files
 
 ```ruby
-dt.optimize.z_order(["a"])
+dt.optimize.z_order(["category"])
 ```
 
 Load a previous version of a table
 
 ```ruby
-dt = DeltaLake::Table.new("./data/delta", version: 1)
+dt = DeltaLake::Table.new("./events", version: 1)
 # or
 dt.load_as_version(1)
 ```

data/ext/deltalake/Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "deltalake"
-version = "0.1.2"
+version = "0.1.3"
 license = "Apache-2.0"
 authors = ["Andrew Kane <andrew@ankane.org>"]
 edition = "2021"
@@ -15,7 +15,7 @@ arrow = { version = "53", features = ["ffi"] }
 arrow-schema = { version = "53", features = ["serde"] }
 chrono = "0.4"
 delta_kernel = "0.4"
-deltalake = { version = "=0.22.2", features = ["azure", "datafusion", "gcs", "s3"] }
+deltalake = { version = "=0.22.3", features = ["azure", "datafusion", "gcs", "s3"] }
 futures = "0.3"
 magnus = "0.7"
 num_cpus = "1"

data/lib/deltalake/table.rb

@@ -188,23 +188,28 @@ module DeltaLake
       JSON.parse(metrics)
     end
 
-    def to_polars(eager: true)
+    def to_polars(eager: true, rechunk: false, columns: nil)
       require "polars-df"
 
       sources = file_uris
-      lf =
-        if sources.empty?
-          Polars::LazyFrame.new
-        else
-          delta_keys = [
-            "AWS_S3_ALLOW_UNSAFE_RENAME",
-            "AWS_S3_LOCKING_PROVIDER",
-            "CONDITIONAL_PUT",
-            "DELTA_DYNAMO_TABLE_NAME"
-          ]
-          storage_options = @storage_options&.reject { |k, _| delta_keys.include?(k.to_s.upcase) }
-          Polars.scan_parquet(sources, storage_options: storage_options)
+      if sources.empty?
+        lf = Polars::LazyFrame.new
+      else
+        delta_keys = [
+          "AWS_S3_ALLOW_UNSAFE_RENAME",
+          "AWS_S3_LOCKING_PROVIDER",
+          "CONDITIONAL_PUT",
+          "DELTA_DYNAMO_TABLE_NAME"
+        ]
+        storage_options = @storage_options&.reject { |k, _| delta_keys.include?(k.to_s.upcase) }
+        lf = Polars.scan_parquet(sources, storage_options: storage_options, rechunk: rechunk)
+
+        if columns
+          # by_name requires polars-df > 0.15.0
+          lf = lf.select(Polars.cs.by_name(*columns))
         end
+      end
+
       eager ? lf.collect : lf
     end
 

data/lib/deltalake/version.rb

@@ -1,3 +1,3 @@
 module DeltaLake
-  VERSION = "0.1.2"
+  VERSION = "0.1.3"
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: deltalake-rb
 version: !ruby/object:Gem::Version
-  version: 0.1.2
+  version: 0.1.3
 platform: ruby
 authors:
 - Andrew Kane
 autorequire:
 bindir: bin
 cert_chain: []
-date: 2024-12-03 00:00:00.000000000 Z
+date: 2024-12-29 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rb_sys