parquet 0.3.3 → 0.4.1

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: b44b6f3ca56a0f4318c361b309c1af213186dcf09e09acbe54561a2dcc920042
-  data.tar.gz: 68aa0636e1467e008ec29bbdb43737904b8f0a26f2d91a5169005f896950687f
+  metadata.gz: 2b5b56cca903ed731d6981d3113e3833a6e6a6a0ffcd301040b32ab0c72bf9c1
+  data.tar.gz: 0e4486e2a67a051852166ac81754c9bfb2807c2ffa51eeda5edb41050432e930
 SHA512:
-  metadata.gz: 30966ae0335b9caa4458a79e5d627b920f2368605785a978690490a7ed05beb1f7ab25eeeed7349cd3c05b49573dc68405cb8c52af5aad5240a42ce4a56b184b
-  data.tar.gz: 5ea6021844baac3bb31acf41f0fdc3795710f9eff21e9c3556692e70f10398249d802b6f3bb22f414d282f2ff33a2ea951fe9d3040625002752d86311eeae160
+  metadata.gz: affe353c972f130973b309ca1ee928928278254830fa39bee8e4bed5452b5b381e2d27f6986067a0f9b27769a78f195887d371dc4825f61096414af92e9edb94
+  data.tar.gz: 9c823757be4b81d3ccafb57571c1d98b530a0d10696712083a0447b4871bf90720ba588d8c48777926b3f7a7f2ffede0c2ed7c9a076420b4dea183b7290ba47e

data/ext/parquet/src/ruby_reader.rs

@@ -184,37 +184,53 @@ impl Length for RubyReader {
             }
             RubyReader::RubyIoLike { inner } => {
                 let unwrapped_inner = ruby.get_inner(*inner);
-                let current_pos = unwrapped_inner.funcall::<_, _, u64>("seek", (0, 1));
 
-                if let Err(e) = current_pos {
+                // Get current position
+                let current_pos = match unwrapped_inner.funcall::<_, _, u64>("pos", ()) {
+                    Ok(pos) => pos,
+                    Err(e) => {
+                        eprintln!("Error seeking: {}", e);
+                        return 0;
+                    }
+                };
+
+                // Seek to end
+                if let Err(e) = unwrapped_inner.funcall::<_, _, u64>("seek", (0, 2)) {
                     eprintln!("Error seeking: {}", e);
                     return 0;
                 }
 
-                if let Err(e) = unwrapped_inner.funcall::<_, _, u64>("seek", (0, 2)) {
+                // Offset at the end of the file is the length of the file
+                let size = match unwrapped_inner.funcall::<_, _, u64>("pos", ()) {
+                    Ok(pos) => pos,
+                    Err(e) => {
+                        eprintln!("Error seeking: {}", e);
+                        return 0;
+                    }
+                };
+
+                // Restore original position
+                if let Err(e) = unwrapped_inner.funcall::<_, _, u64>("seek", (current_pos, 0)) {
                     eprintln!("Error seeking: {}", e);
                     return 0;
                 }
 
-                let size = unwrapped_inner.funcall::<_, _, u64>("pos", ());
-
-                match size {
-                    Ok(size) => {
-                        // Restore original position
-                        if let Err(e) = unwrapped_inner.funcall::<_, _, u64>(
-                            "seek",
-                            (current_pos.expect("Current position is not set!"), 0),
-                        ) {
-                            eprintln!("Error seeking: {}", e);
-                            return 0;
-                        }
-                        size
-                    }
+                let final_pos = match unwrapped_inner.funcall::<_, _, u64>("pos", ()) {
+                    Ok(pos) => pos,
                     Err(e) => {
                         eprintln!("Error seeking: {}", e);
                         return 0;
                     }
-                }
+                };
+
+                assert_eq!(
+                    current_pos, final_pos,
+                    "Failed to restore original position in seekable IO object. Started at position {}, but ended at {}",
+                    current_pos,
+                    final_pos
+                );
+
+                size
             }
         }
     }

data/ext/parquet/src/types/record_types.rs

@@ -136,7 +136,8 @@ impl TryIntoValue for ParquetField {
                         .map_err(|e| ReaderError::Utf8Error(e))
                         .and_then(|s| Ok(s.into_value_with(handle)))?)
                 } else {
-                    Ok(handle.str_from_slice(s.as_bytes()).as_value())
+                    let s = String::from_utf8_lossy(s.as_bytes());
+                    Ok(s.into_value_with(handle))
                 }
             }
             Field::Byte(b) => Ok(b.into_value_with(handle)),

data/ext/parquet/src/utils.rs

@@ -83,7 +83,7 @@ pub fn parse_parquet_rows_args(ruby: &Ruby, args: &[Value]) -> Result<ParquetRow
         None => ParserResultType::Hash,
     };
 
-    let strict = kwargs.optional.2.flatten().unwrap_or(false);
+    let strict = kwargs.optional.2.flatten().unwrap_or(true);
 
     Ok(ParquetRowsArgs {
         to_read,
@@ -159,6 +159,6 @@ pub fn parse_parquet_columns_args(
         result_type,
         columns: kwargs.optional.1.flatten(),
         batch_size: kwargs.optional.2.flatten(),
-        strict: kwargs.optional.3.flatten().unwrap_or(false),
+        strict: kwargs.optional.3.flatten().unwrap_or(true),
     })
 }

data/lib/parquet/version.rb

@@ -1,3 +1,3 @@
 module Parquet
-  VERSION = "0.3.3"
+  VERSION = "0.4.1"
 end

data/lib/parquet.rbi

@@ -11,7 +11,8 @@ module Parquet
     params(
       input: T.any(String, File, StringIO, IO),
       result_type: T.nilable(T.any(String, Symbol)),
-      columns: T.nilable(T::Array[String])
+      columns: T.nilable(T::Array[String]),
+      strict: T.nilable(T::Boolean)
     ).returns(T::Enumerator[T.any(T::Hash[String, T.untyped], T::Array[T.untyped])])
   end
   sig do
@@ -19,10 +20,11 @@ module Parquet
       input: T.any(String, File, StringIO, IO),
       result_type: T.nilable(T.any(String, Symbol)),
       columns: T.nilable(T::Array[String]),
+      strict: T.nilable(T::Boolean),
       blk: T.nilable(T.proc.params(row: T.any(T::Hash[String, T.untyped], T::Array[T.untyped])).void)
     ).returns(NilClass)
   end
-  def self.each_row(input, result_type: nil, columns: nil, &blk)
+  def self.each_row(input, result_type: nil, columns: nil, strict: nil, &blk)
   end
 
   # Options:
@@ -36,7 +38,8 @@ module Parquet
       input: T.any(String, File, StringIO, IO),
       result_type: T.nilable(T.any(String, Symbol)),
       columns: T.nilable(T::Array[String]),
-      batch_size: T.nilable(Integer)
+      batch_size: T.nilable(Integer),
+      strict: T.nilable(T::Boolean)
     ).returns(T::Enumerator[T.any(T::Hash[String, T.untyped], T::Array[T.untyped])])
   end
   sig do
@@ -45,11 +48,12 @@ module Parquet
       result_type: T.nilable(T.any(String, Symbol)),
       columns: T.nilable(T::Array[String]),
       batch_size: T.nilable(Integer),
+      strict: T.nilable(T::Boolean),
       blk:
         T.nilable(T.proc.params(batch: T.any(T::Hash[String, T::Array[T.untyped]], T::Array[T::Array[T.untyped]])).void)
     ).returns(NilClass)
   end
-  def self.each_column(input, result_type: nil, columns: nil, batch_size: nil, &blk)
+  def self.each_column(input, result_type: nil, columns: nil, batch_size: nil, strict: nil, &blk)
   end
 
   # Options:
@@ -80,7 +84,15 @@ module Parquet
       sample_size: T.nilable(Integer)
     ).void
   end
-  def self.write_rows(read_from, schema:, write_to:, batch_size: nil, flush_threshold: nil, compression: nil, sample_size: nil)
+  def self.write_rows(
+    read_from,
+    schema:,
+    write_to:,
+    batch_size: nil,
+    flush_threshold: nil,
+    compression: nil,
+    sample_size: nil
+  )
   end
 
   # Options:

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: parquet
 version: !ruby/object:Gem::Version
-  version: 0.3.3
+  version: 0.4.1
 platform: ruby
 authors:
 - Nathan Jaremko
 autorequire:
 bindir: bin
 cert_chain: []
-date: 2025-02-19 00:00:00.000000000 Z
+date: 2025-02-20 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rb_sys