polars-df 0.9.0 → 0.10.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: d6852d4e91c063866f1a64c32607e62f536c1faa5730c57f278632b41d418614
-  data.tar.gz: 3c3aef0366fb2e182826a6fc70b5b278974cd71ce9a968ea2faf7dd6c3324a8c
+  metadata.gz: 33ce29f4e785d3152582767ea838dc4636c7585b44508b23468cc133d2e00c9a
+  data.tar.gz: 0243b5580f8e85439b6b41e1572e9e630b2d610ebfe647c2f09683f029f3026c
 SHA512:
-  metadata.gz: 7ce8f3a7c5e5f01d673361a5cd070dbdd675099794cebdd12c29dc5e5963b2b7d0798f432f32e917ed5d77eb2ffe9f5def99d9dbd616b04a45e5483d562c49e1
-  data.tar.gz: f69de773948caf4dcf48a59f52c881ae71b1a2a45f16aad65bc8aea0c4b4c31b2603a02907b62718d8f345dcb85ba1ec46abc2d3356cd2603c16a26cac596f60
+  metadata.gz: b55fa78a8d21213151ff6729d950dc62fd17fb14841046dbd2b076be713cbf158df2da0b5f49c5bc86eb8c54e502f33d96b4468b5087948f0f3fa6aec1fee083
+  data.tar.gz: 959d5781a9541f37d614b48a7d2f531fadd39d67b6dcf0f8afa5593a082bc3106366046548d6a6d641ed7c060efd14da915949eb49f85adfa81e6be381148749

data/CHANGELOG.md

@@ -1,3 +1,15 @@
+## 0.10.0 (2024-05-02)
+
+- Updated Polars to 0.39.2
+- Added support for writing JSON to string
+- Added support for writing Parquet to `StringIO`
+- Added support for cross joins
+- Added `data_page_size` option to `write_parquet` method
+- Added `truncate_ragged_lines` option to `read_csv`, `read_csv_batched`, and `scan_csv` methods
+- Added precompiled gem for Linux x86-64 MUSL
+- Changed `drop` method to ignore missing columns
+- Fixed error with `then` method
+
 ## 0.9.0 (2024-03-03)
 
 See the [upgrade guide](https://docs.pola.rs/releases/upgrade/0.20/)

data/Cargo.lock

@@ -83,9 +83,9 @@ dependencies = [
 
 [[package]]
 name = "argminmax"
-version = "0.6.1"
+version = "0.6.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "202108b46429b765ef483f8a24d5c46f48c14acfdacc086dd4ab6dddf6bcdbd2"
+checksum = "52424b59d69d69d5056d508b260553afd91c57e21849579cd1f50ee8b8b88eaa"
 dependencies = [
  "num-traits",
 ]
@@ -300,9 +300,9 @@ checksum = "baf1de4339761588bc0619e3cbc0120ee582ebb74b53b4efbf79117bd2da40fd"
 
 [[package]]
 name = "chrono"
-version = "0.4.31"
+version = "0.4.35"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "7f2c685bad3eb3d45a01354cedb7d5faa66194d1d58ba6e267a8de788f79db38"
+checksum = "8eaf5903dcbc0a39312feb77df2ff4c76387d591b9fc7b04a238dcf8bb62639a"
 dependencies = [
  "android-tzdata",
  "iana-time-zone",
@@ -310,7 +310,7 @@ dependencies = [
  "num-traits",
  "serde",
  "wasm-bindgen",
- "windows-targets 0.48.5",
+ "windows-targets 0.52.0",
 ]
 
 [[package]]
@@ -966,9 +966,9 @@ dependencies = [
 
 [[package]]
 name = "magnus"
-version = "0.6.2"
+version = "0.6.3"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "4778544796676e8428e9c622460ebf284bea52d8b10db3aeb449d8b5e61b3a13"
+checksum = "0fc7a31fb0b64761e3cd09a6975577601fccc5f08b8fc9245064fc4f71ed6a9d"
 dependencies = [
  "magnus-macros",
  "rb-sys",
@@ -1243,7 +1243,7 @@ dependencies = [
 
 [[package]]
 name = "polars"
-version = "0.9.0"
+version = "0.10.0"
 dependencies = [
  "ahash",
  "chrono",
@@ -1251,7 +1251,7 @@ dependencies = [
  "jemallocator",
  "magnus",
  "mimalloc",
- "polars 0.38.1",
+ "polars 0.39.2",
  "polars-core",
  "polars-parquet",
  "polars-utils",
@@ -1261,9 +1261,9 @@ dependencies = [
 
 [[package]]
 name = "polars"
-version = "0.38.1"
+version = "0.39.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "3a26ef94cfedd5915da990a0b4740cca17b5854bd44a8e8c741fe732c02aac37"
+checksum = "0ea21b858b16b9c0e17a12db2800d11aa5b4bd182be6b3022eb537bbfc1f2db5"
 dependencies = [
  "getrandom",
  "polars-arrow",
@@ -1282,9 +1282,9 @@ dependencies = [
 
 [[package]]
 name = "polars-arrow"
-version = "0.38.1"
+version = "0.39.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "48e71d30a9fa503bc3baaff3b4c48f08d402c442a50ea7fb9d475ce7b575425a"
+checksum = "725b09f2b5ef31279b66e27bbab63c58d49d8f6696b66b1f46c7eaab95e80f75"
 dependencies = [
  "ahash",
  "atoi",
@@ -1330,9 +1330,9 @@ dependencies = [
 
 [[package]]
 name = "polars-compute"
-version = "0.38.1"
+version = "0.39.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "26342dea46502e8a3322f484062869c2fa49185d512bce4fb44f350b559b4eae"
+checksum = "a796945b14b14fbb79b91ef0406e6fddca2be636e889f81ea5d6ee7d36efb4fe"
 dependencies = [
  "bytemuck",
  "either",
@@ -1346,9 +1346,9 @@ dependencies = [
 
 [[package]]
 name = "polars-core"
-version = "0.38.1"
+version = "0.39.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "99e0885a8f1bd1f4d928f5eaa852825bf647b6b5e21e171b6af838f77b6565f3"
+checksum = "465f70d3e96b6d0b1a43c358ba451286b8c8bd56696feff020d65702aa33e35c"
 dependencies = [
  "ahash",
  "bitflags 2.4.1",
@@ -1380,9 +1380,9 @@ dependencies = [
 
 [[package]]
 name = "polars-error"
-version = "0.38.1"
+version = "0.39.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "d259d905c17d8e8b2de1eadc94dc4186bf1d325f1be81b4087afea22a6f753d6"
+checksum = "5224d5d05e6b8a6f78b75951ae1b5f82c8ab1979e11ffaf5fd41941e3d5b0757"
 dependencies = [
  "avro-schema",
  "polars-arrow-format",
@@ -1393,9 +1393,9 @@ dependencies = [
 
 [[package]]
 name = "polars-io"
-version = "0.38.1"
+version = "0.39.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "45f694b918ba2ee7e6f13e8415598f94009c390a9e61c95e6b9c26c8fe1a1a54"
+checksum = "b2c8589e418cbe4a48228d64b2a8a40284a82ec3c98817c0c2bcc0267701338b"
 dependencies = [
  "ahash",
  "async-trait",
@@ -1433,9 +1433,9 @@ dependencies = [
 
 [[package]]
 name = "polars-json"
-version = "0.38.1"
+version = "0.39.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "fd7c4d33540a22cd49ccb9b4da77999604ba986707d4568be57fde2f719954ae"
+checksum = "81224492a649a12b668480c0cf219d703f432509765d2717e72fe32ad16fc701"
 dependencies = [
  "ahash",
  "chrono",
@@ -1454,9 +1454,9 @@ dependencies = [
 
 [[package]]
 name = "polars-lazy"
-version = "0.38.1"
+version = "0.39.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "56e3b40272d24142bcecb2979b19ec8d8c1a14036cb3cea09ce8fb8a4a43bcde"
+checksum = "89b2632b1af668e2058d5f8f916d8fbde3cac63d03ae29a705f598e41dcfeb7f"
 dependencies = [
  "ahash",
  "bitflags 2.4.1",
@@ -1478,9 +1478,9 @@ dependencies = [
 
 [[package]]
 name = "polars-ops"
-version = "0.38.1"
+version = "0.39.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "cd5bad61c2fa1977eb65bb719f12d4f68b908edf1106b91b3ab9615f9df8843e"
+checksum = "efdbdb4d9a92109bc2e0ce8e17af5ae8ab643bb5b7ee9d1d74f0aeffd1fbc95f"
 dependencies = [
  "ahash",
  "aho-corasick",
@@ -1515,9 +1515,9 @@ dependencies = [
 
 [[package]]
 name = "polars-parquet"
-version = "0.38.1"
+version = "0.39.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "06d84fb9b005a19ca523406df371d9329466ae87df48922d0d3d8955072502a4"
+checksum = "b421d2196f786fdfe162db614c8485f8308fe41575d4de634a39bbe460d1eb6a"
 dependencies = [
  "ahash",
  "async-stream",
@@ -1541,9 +1541,9 @@ dependencies = [
 
 [[package]]
 name = "polars-pipe"
-version = "0.38.1"
+version = "0.39.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "58097bef7208a5b833c4d832d948026854917b4a219d55ab1779eb36b59fac0f"
+checksum = "48700f1d5bd56a15451e581f465c09541492750360f18637b196f995470a015c"
 dependencies = [
  "crossbeam-channel",
  "crossbeam-queue",
@@ -1566,14 +1566,15 @@ dependencies = [
 
 [[package]]
 name = "polars-plan"
-version = "0.38.1"
+version = "0.39.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "56493c0e13aaccfcae59985db34da30cd4893e57edc9715d8688c96d7e911d47"
+checksum = "2fb8e2302e20c44defd5be8cad9c96e75face63c3a5f609aced8c4ec3b3ac97d"
 dependencies = [
  "ahash",
  "bytemuck",
  "chrono",
  "chrono-tz",
+ "hashbrown 0.14.3",
  "once_cell",
  "percent-encoding",
  "polars-arrow",
@@ -1585,6 +1586,7 @@ dependencies = [
  "polars-time",
  "polars-utils",
  "rayon",
+ "recursive",
  "regex",
  "serde",
  "smartstring",
@@ -1594,9 +1596,9 @@ dependencies = [
 
 [[package]]
 name = "polars-row"
-version = "0.38.1"
+version = "0.39.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "7def6f9fc14fbfc0550bad615a757f3e1d86c00983c5ff23166fcdf205438d51"
+checksum = "a515bdc68c2ae3702e3de70d89601f3b71ca8137e282a226dddb53ee4bacfa2e"
 dependencies = [
  "bytemuck",
  "polars-arrow",
@@ -1606,9 +1608,9 @@ dependencies = [
 
 [[package]]
 name = "polars-sql"
-version = "0.38.1"
+version = "0.39.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "4f9d7de9dca8170a20b6c4cb7bafaf724abe88e807646bc3c2e98f13a34a7c4c"
+checksum = "7b4bb7cc1c04c3023d1953b2f1dec50515e8fd8169a5a2bf4967b3b082232db7"
 dependencies = [
  "hex",
  "polars-arrow",
@@ -1624,9 +1626,9 @@ dependencies = [
 
 [[package]]
 name = "polars-time"
-version = "0.38.1"
+version = "0.39.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "162c815c3cb0f859da40f056c8a0a9c4247900e1275702ae399192ea60acac2a"
+checksum = "efc18e3ad92eec55db89d88f16c22d436559ba7030cf76f86f6ed7a754b673f1"
 dependencies = [
  "atoi",
  "chrono",
@@ -1645,9 +1647,9 @@ dependencies = [
 
 [[package]]
 name = "polars-utils"
-version = "0.38.1"
+version = "0.39.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "7fca7938ee789314ac92a0bf6c1c4e5eaeb5e428241df2519fd70f21dba49194"
+checksum = "c760b6c698cfe2fbbbd93d6cfb408db14ececfe1d92445dae2229ce1b5b21ae8"
 dependencies = [
  "ahash",
  "bytemuck",
@@ -1659,6 +1661,7 @@ dependencies = [
  "raw-cpuid",
  "rayon",
  "smartstring",
+ "stacker",
  "sysinfo",
  "version_check",
 ]
@@ -1678,6 +1681,15 @@ dependencies = [
  "unicode-ident",
 ]
 
+[[package]]
+name = "psm"
+version = "0.1.21"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "5787f7cda34e3033a72192c018bc5883100330f362ef279a8cbccfce8bb4e874"
+dependencies = [
+ "cc",
+]
+
 [[package]]
 name = "quote"
 version = "1.0.35"
@@ -1758,18 +1770,18 @@ dependencies = [
 
 [[package]]
 name = "rb-sys"
-version = "0.9.90"
+version = "0.9.97"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "55d933382388cc7a6fdfd54e222eca7994791ac4b9ce5c9e8df280c739d86bbe"
+checksum = "47d30bcad206b51f2f66121190ca678dce1fdf3a2eae0ac5d838d1818b19bdf5"
 dependencies = [
  "rb-sys-build",
 ]
 
 [[package]]
 name = "rb-sys-build"
-version = "0.9.90"
+version = "0.9.97"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "ebc5a7e3a875419baaa0d8cc606cdfb9361b444cb7e5abcf0de4693025887374"
+checksum = "3cbd92f281615f3c2dcb9dcb0f0576624752afbf9a7f99173b37c4b55b62dd8a"
 dependencies = [
  "bindgen",
  "lazy_static",
@@ -1786,6 +1798,26 @@ version = "0.1.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "a35802679f07360454b418a5d1735c89716bde01d35b1560fc953c1415a0b3bb"
 
+[[package]]
+name = "recursive"
+version = "0.1.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "0786a43debb760f491b1bc0269fe5e84155353c67482b9e60d0cfb596054b43e"
+dependencies = [
+ "recursive-proc-macro-impl",
+ "stacker",
+]
+
+[[package]]
+name = "recursive-proc-macro-impl"
+version = "0.1.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "76009fbe0614077fc1a2ce255e3a1881a2e3a3527097d5dc6d8212c585e7e38b"
+dependencies = [
+ "quote",
+ "syn 2.0.46",
+]
+
 [[package]]
 name = "redox_syscall"
 version = "0.4.1"
@@ -2012,6 +2044,19 @@ dependencies = [
  "log",
 ]
 
+[[package]]
+name = "stacker"
+version = "0.1.15"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "c886bd4480155fd3ef527d45e9ac8dd7118a898a46530b7b94c3e21866259fce"
+dependencies = [
+ "cc",
+ "cfg-if",
+ "libc",
+ "psm",
+ "winapi",
+]
+
 [[package]]
 name = "static_assertions"
 version = "1.1.0"

data/README.md

@@ -433,4 +433,5 @@ cd polars-ruby
 bundle install
 bundle exec rake compile
 bundle exec rake test
+bundle exec rake test:docs
 ```

data/ext/polars/Cargo.toml

@@ -1,10 +1,10 @@
 [package]
 name = "polars"
-version = "0.9.0"
+version = "0.10.0"
 license = "MIT"
 authors = ["Andrew Kane <andrew@ankane.org>"]
 edition = "2021"
-rust-version = "1.73.0"
+rust-version = "1.76.0"
 publish = false
 
 [lib]
@@ -15,14 +15,14 @@ ahash = "0.8"
 chrono = "0.4"
 either = "1.8"
 magnus = "0.6"
-polars-core = "=0.38.1"
-polars-parquet = "=0.38.1"
-polars-utils = "=0.38.1"
+polars-core = "=0.39.2"
+polars-parquet = "=0.39.2"
+polars-utils = "=0.39.2"
 serde_json = "1"
 smartstring = "1"
 
 [dependencies.polars]
-version = "=0.38.1"
+version = "=0.39.2"
 features = [
     "abs",
     "approx_unique",
@@ -34,6 +34,7 @@ features = [
     "binary_encoding",
     "concat_str",
     "cov",
+    "cross_join",
     "cse",
     "csv",
     "cum_agg",
@@ -47,6 +48,7 @@ features = [
     "dtype-full",
     "dynamic_group_by",
     "ewma",
+    "ewma_by",
     "extract_groups",
     "extract_jsonpath",
     "find_many",

data/ext/polars/src/batched_csv.rs

@@ -49,6 +49,7 @@ impl RbBatchedCsv {
         let row_index = Option::<(String, IdxSize)>::try_convert(arguments[21])?;
         let sample_size = usize::try_convert(arguments[22])?;
         let eol_char = String::try_convert(arguments[23])?;
+        let truncate_ragged_lines = bool::try_convert(arguments[24])?;
         // end arguments
 
         let null_values = null_values.map(|w| w.0);
@@ -107,7 +108,8 @@ impl RbBatchedCsv {
             .with_end_of_line_char(eol_char)
             .with_skip_rows_after_header(skip_rows_after_header)
             .with_row_index(row_index)
-            .sample_size(sample_size);
+            .sample_size(sample_size)
+            .truncate_ragged_lines(truncate_ragged_lines);
 
         let reader = if low_memory {
             let reader = reader

data/ext/polars/src/conversion/anyvalue.rs

@@ -3,8 +3,8 @@ use magnus::{
     class, prelude::*, r_hash::ForEach, Float, Integer, IntoValue, RArray, RHash, RString, Ruby,
     TryConvert, Value,
 };
-use polars::frame::row::any_values_to_dtype;
 use polars::prelude::*;
+use polars_core::utils::any_values_to_supertype_and_n_dtypes;
 
 use super::{struct_dict, ObjectValue, Wrap};
 
@@ -120,7 +120,8 @@ impl<'s> TryConvert for Wrap<AnyValue<'s>> {
                     avs.push(Wrap::<AnyValue>::try_convert(item?)?.0)
                 }
 
-                let (dtype, _n_types) = any_values_to_dtype(&avs).map_err(RbPolarsErr::from)?;
+                let (dtype, _n_types) =
+                    any_values_to_supertype_and_n_dtypes(&avs).map_err(RbPolarsErr::from)?;
 
                 // push the rest
                 avs.reserve(list.len());

data/ext/polars/src/conversion/mod.rs

@@ -217,10 +217,7 @@ impl IntoValue for Wrap<DataType> {
                 let class = pl.const_get::<_, Value>("Enum").unwrap();
                 let s = Series::from_arrow("category", categories.to_boxed()).unwrap();
                 let series = to_series(s.into());
-                class
-                    .funcall::<_, _, Value>("new", (series,))
-                    .unwrap()
-                    .into()
+                class.funcall::<_, _, Value>("new", (series,)).unwrap()
             }
             DataType::Time => {
                 let class = pl.const_get::<_, Value>("Time").unwrap();
@@ -740,8 +737,7 @@ impl TryConvert for Wrap<JoinType> {
             "outer_coalesce" => JoinType::Outer { coalesce: true },
             "semi" => JoinType::Semi,
             "anti" => JoinType::Anti,
-            // #[cfg(feature = "cross_join")]
-            // "cross" => JoinType::Cross,
+            "cross" => JoinType::Cross,
             v => {
                 return Err(RbValueError::new_err(format!(
                 "how must be one of {{'inner', 'left', 'outer', 'semi', 'anti', 'cross'}}, got {}",
@@ -785,6 +781,21 @@ impl TryConvert for Wrap<ListToStructWidthStrategy> {
     }
 }
 
+impl TryConvert for Wrap<NonExistent> {
+    fn try_convert(ob: Value) -> RbResult<Self> {
+        let parsed = match String::try_convert(ob)?.as_str() {
+            "null" => NonExistent::Null,
+            "raise" => NonExistent::Raise,
+            v => {
+                return Err(RbValueError::new_err(format!(
+                    "`non_existent` must be one of {{'null', 'raise'}}, got {v}",
+                )))
+            }
+        };
+        Ok(Wrap(parsed))
+    }
+}
+
 impl TryConvert for Wrap<NullBehavior> {
     fn try_convert(ob: Value) -> RbResult<Self> {
         let parsed = match String::try_convert(ob)?.as_str() {
@@ -1066,7 +1077,7 @@ impl TryConvert for Wrap<NonZeroUsize> {
     fn try_convert(ob: Value) -> RbResult<Self> {
         let v = usize::try_convert(ob)?;
         NonZeroUsize::new(v)
-            .map(|v| Wrap(v))
+            .map(Wrap)
             .ok_or(RbValueError::new_err("must be non-zero".into()))
     }
 }

data/ext/polars/src/dataframe.rs

@@ -16,7 +16,7 @@ use std::num::NonZeroUsize;
 use std::ops::Deref;
 
 use crate::conversion::*;
-use crate::file::{get_file_like, get_mmap_bytes_reader};
+use crate::file::{get_either_file, get_file_like, get_mmap_bytes_reader, EitherRustRubyFile};
 use crate::map::dataframe::{
     apply_lambda_unknown, apply_lambda_with_bool_out_type, apply_lambda_with_primitive_out_type,
     apply_lambda_with_utf8_out_type,
@@ -136,6 +136,7 @@ impl RbDataFrame {
         let row_index = Option::<(String, IdxSize)>::try_convert(arguments[22])?;
         let sample_size = usize::try_convert(arguments[23])?;
         let eol_char = String::try_convert(arguments[24])?;
+        let truncate_ragged_lines = bool::try_convert(arguments[25])?;
         // end arguments
 
         let null_values = null_values.map(|w| w.0);
@@ -196,6 +197,7 @@ impl RbDataFrame {
             .with_skip_rows_after_header(skip_rows_after_header)
             .with_row_index(row_index)
             .sample_size(sample_size)
+            .truncate_ragged_lines(truncate_ragged_lines)
             .finish()
             .map_err(RbPolarsErr::from)?;
         Ok(df.into())
@@ -213,19 +215,34 @@ impl RbDataFrame {
         use_statistics: bool,
         rechunk: bool,
     ) -> RbResult<Self> {
+        use EitherRustRubyFile::*;
+
         let row_index = row_index.map(|(name, offset)| RowIndex { name, offset });
-        let mmap_bytes_r = get_mmap_bytes_reader(rb_f)?;
-        let df = ParquetReader::new(mmap_bytes_r)
-            .with_projection(projection)
-            .with_columns(columns)
-            .read_parallel(parallel.0)
-            .with_n_rows(n_rows)
-            .with_row_index(row_index)
-            .set_low_memory(low_memory)
-            .use_statistics(use_statistics)
-            .set_rechunk(rechunk)
-            .finish()
-            .map_err(RbPolarsErr::from)?;
+        let result = match get_either_file(rb_f, false)? {
+            Rb(f) => {
+                let buf = f.as_buffer();
+                ParquetReader::new(buf)
+                    .with_projection(projection)
+                    .with_columns(columns)
+                    .read_parallel(parallel.0)
+                    .with_n_rows(n_rows)
+                    .with_row_index(row_index)
+                    .set_low_memory(low_memory)
+                    .use_statistics(use_statistics)
+                    .set_rechunk(rechunk)
+                    .finish()
+            }
+            Rust(f) => ParquetReader::new(f.into_inner())
+                .with_projection(projection)
+                .with_columns(columns)
+                .read_parallel(parallel.0)
+                .with_n_rows(n_rows)
+                .with_row_index(row_index)
+                .use_statistics(use_statistics)
+                .set_rechunk(rechunk)
+                .finish(),
+        };
+        let df = result.map_err(RbPolarsErr::from)?;
         Ok(RbDataFrame::new(df))
     }
 
@@ -576,6 +593,7 @@ impl RbDataFrame {
         compression_level: Option<i32>,
         statistics: bool,
         row_group_size: Option<usize>,
+        data_page_size: Option<usize>,
     ) -> RbResult<()> {
         let compression = parse_parquet_compression(&compression, compression_level)?;
 
@@ -585,10 +603,18 @@ impl RbDataFrame {
                 .with_compression(compression)
                 .with_statistics(statistics)
                 .with_row_group_size(row_group_size)
+                .with_data_page_size(data_page_size)
                 .finish(&mut self.df.borrow_mut())
                 .map_err(RbPolarsErr::from)?;
         } else {
-            todo!();
+            let buf = get_file_like(rb_f, true)?;
+            ParquetWriter::new(buf)
+                .with_compression(compression)
+                .with_statistics(statistics)
+                .with_row_group_size(row_group_size)
+                .with_data_page_size(data_page_size)
+                .finish(&mut self.df.borrow_mut())
+                .map_err(RbPolarsErr::from)?;
         }
 
         Ok(())

data/ext/polars/src/expr/array.rs

@@ -59,8 +59,12 @@ impl RbExpr {
         self.inner.clone().arr().arg_max().into()
     }
 
-    pub fn arr_get(&self, index: &RbExpr) -> Self {
-        self.inner.clone().arr().get(index.inner.clone()).into()
+    pub fn arr_get(&self, index: &RbExpr, null_on_oob: bool) -> Self {
+        self.inner
+            .clone()
+            .arr()
+            .get(index.inner.clone(), null_on_oob)
+            .into()
     }
 
     pub fn arr_join(&self, separator: &RbExpr, ignore_nulls: bool) -> Self {

data/ext/polars/src/expr/datetime.rs

@@ -37,11 +37,16 @@ impl RbExpr {
         self.inner.clone().dt().cast_time_unit(tu.0).into()
     }
 
-    pub fn dt_replace_time_zone(&self, time_zone: Option<String>, ambiguous: &Self) -> Self {
+    pub fn dt_replace_time_zone(
+        &self,
+        time_zone: Option<String>,
+        ambiguous: &Self,
+        non_existent: Wrap<NonExistent>,
+    ) -> Self {
         self.inner
             .clone()
             .dt()
-            .replace_time_zone(time_zone, ambiguous.inner.clone())
+            .replace_time_zone(time_zone, ambiguous.inner.clone(), non_existent.0)
             .into()
     }