Port re_format_arrow to arrow-rs (#8664)

* Part of #3741 The "Print Datastore" feature will produce a slightly different result.
rerun-io · Jan 14, 2025 · e4f8df4 · e4f8df4
1 parent 1ffe586
commit e4f8df4
Show file tree

Hide file tree

Showing 5 changed files with 125 additions and 118 deletions.
diff --git a/Cargo.lock b/Cargo.lock
@@ -5980,9 +5980,9 @@ dependencies = [
 name = "re_format_arrow"
 version = "0.22.0-alpha.1+dev"
 dependencies = [
+ "arrow",
  "comfy-table",
  "itertools 0.13.0",
- "re_arrow2",
  "re_tuid",
  "re_types_core",
 ]

diff --git a/crates/store/re_chunk/src/transport.rs b/crates/store/re_chunk/src/transport.rs
@@ -1,4 +1,6 @@
-use arrow::array::{ArrayRef as ArrowArrayRef, StructArray as ArrowStructArray};
+use arrow::array::{
+    Array as ArrowArray, ArrayRef as ArrowArrayRef, StructArray as ArrowStructArray,
+};
 use arrow2::{
     array::{Array as Arrow2Array, ListArray},
     chunk::Chunk as Arrow2Chunk,
@@ -48,10 +50,20 @@ pub struct TransportChunk {
 impl std::fmt::Display for TransportChunk {
     #[inline]
     fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
+        // TODO(#3741): simplify code when we have migrated to arrow-rs
         re_format_arrow::format_dataframe(
-            &self.schema.metadata,
-            &self.schema.fields,
-            self.data.iter().map(|list_array| &**list_array),
+            &self.schema.metadata.clone().into_iter().collect(),
+            &self
+                .schema
+                .fields
+                .iter()
+                .map(|field| arrow::datatypes::Field::from(field.clone()))
+                .collect(),
+            &self
+                .data
+                .iter()
+                .map(|list_array| ArrowArrayRef::from(list_array.clone()))
+                .collect_vec(),
         )
         .fmt(f)
     }

diff --git a/crates/store/re_chunk_store/tests/snapshots/formatting__format_chunk_store.snap b/crates/store/re_chunk_store/tests/snapshots/formatting__format_chunk_store.snap
@@ -24,11 +24,11 @@ ChunkStore {
         │ ┌──────────────────────────────────┬───────────────┬───────────────────────────────┬───────────────────┬───────────────────┐ │
         │ │ RowId                            ┆ frame_nr      ┆ log_time                      ┆ example.MyColor   ┆ example.MyIndex   │ │
         │ │ ---                              ┆ ---           ┆ ---                           ┆ ---               ┆ ---               │ │
-        │ │ type: "struct[2]"                ┆ type: "i64"   ┆ type: "timestamp(ns)"         ┆ type: "list[u32]" ┆ type: "list[u64]" │ │
+        │ │ type: "Struct[2]"                ┆ type: "i64"   ┆ type: "Timestamp(ns)"         ┆ type: "List[u32]" ┆ type: "List[u64]" │ │
         │ │ ARROW:extension:name: "TUID"     ┆ is_sorted: "" ┆ is_sorted: ""                 ┆ kind: "data"      ┆ kind: "data"      │ │
         │ │ kind: "control"                  ┆ kind: "time"  ┆ kind: "time"                  ┆                   ┆                   │ │
         │ ╞══════════════════════════════════╪═══════════════╪═══════════════════════════════╪═══════════════════╪═══════════════════╡ │
-        │ │ 0000000067816A6BB4B8C1254D40007B ┆ 1             ┆ 2025-01-10 18:43:42.123456789 ┆ [0, 1, 2]         ┆ [0, 1, 2]         │ │
+        │ │ 0000000067816A6BB4B8C1254D40007B ┆ 1             ┆ 2025-01-10T18:43:42.123456789 ┆ [0, 1, 2]         ┆ [0, 1, 2]         │ │
         │ └──────────────────────────────────┴───────────────┴───────────────────────────────┴───────────────────┴───────────────────┘ │
         └──────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────┘
     ]

diff --git a/crates/store/re_format_arrow/Cargo.toml b/crates/store/re_format_arrow/Cargo.toml
@@ -20,7 +20,7 @@ all-features = true
 
 
 [dependencies]
-arrow2.workspace = true
+arrow.workspace = true
 itertools.workspace = true
 re_tuid.workspace = true
 re_types_core.workspace = true # tuid serialization