Impl custom PartialEq and Hash on Field that ignores metadata

Eventual-Inc · Feb 24, 2024 · 079991f · 079991f
1 parent 412ffbc
commit 079991f
Show file tree

Hide file tree

Showing 3 changed files with 17 additions and 9 deletions.
diff --git a/src/daft-core/src/datatypes/field.rs b/src/daft-core/src/datatypes/field.rs
@@ -10,7 +10,7 @@ use serde::{Deserialize, Serialize};
 
 pub type Metadata = std::collections::BTreeMap<String, String>;
 
-#[derive(Clone, Debug, PartialEq, Eq, Deserialize, Serialize, Hash)]
+#[derive(Clone, Debug, Eq, Deserialize, Serialize)]
 pub struct Field {
     pub name: String,
     pub dtype: DataType,
@@ -129,3 +129,16 @@ impl Display for Field {
         write!(f, "{}#{}", self.name, self.dtype)
     }
 }
+
+impl PartialEq for Field {
+    fn eq(&self, other: &Self) -> bool {
+        self.dtype == other.dtype && self.name == other.name
+    }
+}
+
+impl std::hash::Hash for Field {
+    fn hash<H: std::hash::Hasher>(&self, state: &mut H) {
+        self.name.hash(state);
+        self.dtype.hash(state);
+    }
+}
diff --git a/src/daft-micropartition/src/micropartition.rs b/src/daft-micropartition/src/micropartition.rs
@@ -354,13 +354,8 @@ impl MicroPartition {
         // Check and validate invariants with asserts
         for table in tables.iter() {
             assert!(
-                table.schema.fields.len() == schema.fields.len()
-                    && table.schema.fields.iter().zip(schema.fields.iter()).all(
-                        |((s1, f1), (s2, f2))| s1 == s2
-                            && f1.name == f2.name
-                            && f1.dtype == f2.dtype
-                    ),
-                "Loaded MicroPartition's tables' schema must match its own schema exactly"
+                table.schema == schema,
+                "Loaded MicroPartition's tables' schema must match its own schema exactly",
             );
         }
 

diff --git a/src/daft-table/src/lib.rs b/src/daft-table/src/lib.rs
@@ -43,7 +43,7 @@ impl Table {
         let mut num_rows = 1;
 
         for (field, series) in schema.fields.values().zip(columns.iter()) {
-            if field.name != series.field().name || field.dtype != series.field().dtype {
+            if field != series.field() {
                 return Err(DaftError::SchemaMismatch(format!("While building a Table, we found that the Schema Field and the Series Field  did not match. schema field: {field} vs series field: {}", series.field())));
             }
             if (series.len() != 1) && (series.len() != num_rows) {