iceberg/scan/
task.rs

1// Licensed to the Apache Software Foundation (ASF) under one
2// or more contributor license agreements.  See the NOTICE file
3// distributed with this work for additional information
4// regarding copyright ownership.  The ASF licenses this file
5// to you under the Apache License, Version 2.0 (the
6// "License"); you may not use this file except in compliance
7// with the License.  You may obtain a copy of the License at
8//
9//   http://www.apache.org/licenses/LICENSE-2.0
10//
11// Unless required by applicable law or agreed to in writing,
12// software distributed under the License is distributed on an
13// "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
14// KIND, either express or implied.  See the License for the
15// specific language governing permissions and limitations
16// under the License.
17
18use std::sync::Arc;
19
20use futures::stream::BoxStream;
21use serde::{Deserialize, Serialize};
22use typed_builder::TypedBuilder;
23
24use crate::expr::BoundPredicate;
25use crate::spec::{
26    DataContentType, DataFileFormat, ManifestEntryRef, NameMapping, PartitionSpec, Schema,
27    SchemaRef, Struct, StructType,
28};
29use crate::{Error, ErrorKind, Result};
30
31/// A stream of [`FileScanTask`].
32pub type FileScanTaskStream = BoxStream<'static, Result<FileScanTask>>;
33
34/// A task to scan part of file.
35#[derive(Debug, Clone, Deserialize, PartialEq, TypedBuilder)]
36#[serde(try_from = "crate::scan::task::_serde::FileScanTaskSerde")]
37#[builder(
38    field_defaults(setter(prefix = "with_")),
39    build_method(into = Result<FileScanTask>)
40)]
41pub struct FileScanTask {
42    /// The total size of the data file in bytes, from the manifest entry.
43    /// Used to skip a stat/HEAD request when reading Parquet footers.
44    file_size_in_bytes: u64,
45    /// The start offset of the file to scan.
46    start: u64,
47    /// The length of the file to scan.
48    length: u64,
49    /// The number of records in the file to scan.
50    ///
51    /// This is an optional field, and only available if we are
52    /// reading the entire data file.
53    #[builder(default)]
54    record_count: Option<u64>,
55
56    /// The first row id assigned to the data file.
57    ///
58    /// Used to derive the `_row_id` metadata column: for a row without an
59    /// explicit `_row_id`, it is this value plus the row's ordinal position.
60    #[builder(default)]
61    first_row_id: Option<i64>,
62
63    /// The data sequence number of the file, as opposed to its file sequence
64    /// number: the sequence number preserved when a file is carried forward
65    /// across a rewrite. May be null for an existing entry in a malformed
66    /// manifest that lacks one.
67    ///
68    /// Used to derive the `_last_updated_sequence_number` metadata column.
69    #[builder(default)]
70    data_sequence_number: Option<i64>,
71
72    /// The data file path corresponding to the task.
73    data_file_path: String,
74
75    /// The format of the file to scan.
76    data_file_format: DataFileFormat,
77
78    /// The schema of the file to scan.
79    schema: SchemaRef,
80    /// The field ids to project.
81    project_field_ids: Vec<i32>,
82    /// The predicate to filter.
83    #[builder(default)]
84    predicate: Option<BoundPredicate>,
85
86    /// The list of delete files that may need to be applied to this data file
87    #[builder(default)]
88    deletes: Vec<FileScanTaskDeleteFile>,
89
90    /// Partition data from the manifest entry, used to identify which columns can use
91    /// constant values from partition metadata vs. reading from the data file.
92    /// Per the Iceberg spec, only identity-transformed partition fields should use constants.
93    #[builder(default)]
94    partition: Option<Struct>,
95
96    /// The partition spec for this file, used to distinguish identity transforms
97    /// (which use partition metadata constants) from non-identity transforms like
98    /// bucket/truncate (which must read source columns from the data file).
99    #[builder(default)]
100    partition_spec: Option<Arc<PartitionSpec>>,
101
102    /// Name mapping from table metadata (property: schema.name-mapping.default),
103    /// used to resolve field IDs from column names when Parquet files lack field IDs
104    /// or have field ID conflicts.
105    #[builder(default)]
106    name_mapping: Option<Arc<NameMapping>>,
107
108    /// The unified partition type across all specs in the table.
109    /// When `RESERVED_FIELD_ID_PARTITION` is in the projected field IDs, the reader
110    /// uses this type along with the task's partition_spec and partition data to
111    /// materialize the `_partition` struct column at read time.
112    ///
113    /// This is a table-level value (same for all tasks in a scan), stored per-task
114    /// so that readers are self-contained without needing back-pointers to table
115    /// metadata. The cost is one Arc clone per task.
116    #[builder(default)]
117    unified_partition_type: Option<Arc<StructType>>,
118
119    /// Whether this scan task should treat column names as case-sensitive when binding predicates.
120    case_sensitive: bool,
121
122    /// Key metadata for encrypted data files (Parquet Modular Encryption).
123    /// When present, the reader uses this to build `FileDecryptionProperties`.
124    ///
125    /// Note on the trust boundary: for the standard encryption scheme this
126    /// carries `StandardKeyMetadata`, whose payload is the *plaintext* DEK.
127    /// Because `FileScanTask` implements [`Serialize`], that plaintext DEK is part
128    /// of the serialized scan plan should these tasks ever be serialized and sent
129    /// over the network.
130    #[builder(default)]
131    key_metadata: Option<Box<[u8]>>,
132}
133
134impl FileScanTask {
135    /// Returns the total size of the data file in bytes.
136    pub fn file_size_in_bytes(&self) -> u64 {
137        self.file_size_in_bytes
138    }
139
140    /// Returns the start offset of the file to scan.
141    pub fn start(&self) -> u64 {
142        self.start
143    }
144
145    /// Returns the length of the file to scan.
146    pub fn length(&self) -> u64 {
147        self.length
148    }
149
150    /// Returns the number of records in the file when the whole file is scanned.
151    pub fn record_count(&self) -> Option<u64> {
152        self.record_count
153    }
154
155    /// Returns the first row id assigned to the data file.
156    pub fn first_row_id(&self) -> Option<i64> {
157        self.first_row_id
158    }
159
160    /// Returns the data sequence number of the file.
161    pub fn data_sequence_number(&self) -> Option<i64> {
162        self.data_sequence_number
163    }
164
165    /// Returns the data file path of this file scan task.
166    pub fn data_file_path(&self) -> &str {
167        &self.data_file_path
168    }
169
170    /// Returns the format of the data file.
171    pub fn data_file_format(&self) -> DataFileFormat {
172        self.data_file_format
173    }
174
175    /// Returns the schema of this file scan task as a reference.
176    pub fn schema(&self) -> &Schema {
177        &self.schema
178    }
179
180    /// Returns the schema of this file scan task as a [`SchemaRef`].
181    pub fn schema_ref(&self) -> SchemaRef {
182        self.schema.clone()
183    }
184
185    /// Returns the project field id of this file scan task.
186    pub fn project_field_ids(&self) -> &[i32] {
187        &self.project_field_ids
188    }
189
190    /// Returns the predicate of this file scan task.
191    pub fn predicate(&self) -> Option<&BoundPredicate> {
192        self.predicate.as_ref()
193    }
194
195    /// Returns the delete files that may need to be applied to the data file.
196    pub fn deletes(&self) -> &[FileScanTaskDeleteFile] {
197        &self.deletes
198    }
199
200    /// Returns the partition data from the manifest entry.
201    pub fn partition(&self) -> Option<&Struct> {
202        self.partition.as_ref()
203    }
204
205    /// Returns the partition spec for the data file.
206    pub fn partition_spec(&self) -> Option<&Arc<PartitionSpec>> {
207        self.partition_spec.as_ref()
208    }
209
210    /// Returns the name mapping used to resolve field ids.
211    pub fn name_mapping(&self) -> Option<&Arc<NameMapping>> {
212        self.name_mapping.as_ref()
213    }
214
215    /// Returns the unified partition type across all table partition specs.
216    pub fn unified_partition_type(&self) -> Option<&Arc<StructType>> {
217        self.unified_partition_type.as_ref()
218    }
219
220    /// Returns whether names are treated as case-sensitive.
221    pub fn case_sensitive(&self) -> bool {
222        self.case_sensitive
223    }
224
225    /// Returns the key metadata for the encrypted data file.
226    pub fn key_metadata(&self) -> Option<&[u8]> {
227        self.key_metadata.as_deref()
228    }
229
230    fn validate(&self) -> Result<()> {
231        match (self.partition.as_ref(), self.partition_spec.as_deref()) {
232            (None, None) => Ok(()),
233            (None, Some(partition_spec)) if partition_spec.is_unpartitioned() => Ok(()),
234            (None, Some(_)) => Err(Error::new(
235                ErrorKind::DataInvalid,
236                "FileScanTask with a partitioned spec requires partition values",
237            )),
238            (Some(partition), None) if partition.fields().is_empty() => Ok(()),
239            (Some(_), None) => Err(Error::new(
240                ErrorKind::DataInvalid,
241                "Non-empty FileScanTask partition requires a partition spec",
242            )),
243            (Some(partition), Some(partition_spec))
244                if partition.fields().len() != partition_spec.fields().len() =>
245            {
246                Err(Error::new(
247                    ErrorKind::DataInvalid,
248                    format!(
249                        "FileScanTask partition has {} fields but partition spec has {} fields",
250                        partition.fields().len(),
251                        partition_spec.fields().len()
252                    ),
253                ))
254            }
255            (Some(_), Some(partition_spec)) => {
256                partition_spec.partition_type(&self.schema)?;
257                Ok(())
258            }
259        }
260    }
261}
262
263impl From<FileScanTask> for Result<FileScanTask> {
264    fn from(task: FileScanTask) -> Self {
265        task.validate()?;
266        Ok(task)
267    }
268}
269
270#[derive(Debug)]
271pub(crate) struct DeleteFileContext {
272    pub(crate) manifest_entry: ManifestEntryRef,
273    pub(crate) partition_spec_id: i32,
274}
275
276impl From<&DeleteFileContext> for FileScanTaskDeleteFile {
277    fn from(ctx: &DeleteFileContext) -> Self {
278        FileScanTaskDeleteFile::builder()
279            .with_file_path(ctx.manifest_entry.file_path().to_string())
280            .with_file_size_in_bytes(ctx.manifest_entry.file_size_in_bytes())
281            .with_file_type(ctx.manifest_entry.content_type())
282            .with_file_format(ctx.manifest_entry.data_file().file_format())
283            .with_partition_spec_id(ctx.partition_spec_id)
284            .with_equality_ids(ctx.manifest_entry.data_file.equality_ids.clone())
285            .with_referenced_data_file(ctx.manifest_entry.data_file.referenced_data_file.clone())
286            .with_content_offset(ctx.manifest_entry.data_file.content_offset)
287            .with_content_size_in_bytes(ctx.manifest_entry.data_file.content_size_in_bytes)
288            .with_record_count(Some(ctx.manifest_entry.record_count()))
289            .with_key_metadata(
290                ctx.manifest_entry
291                    .data_file
292                    .key_metadata
293                    .as_deref()
294                    .map(Box::from),
295            )
296            .build()
297    }
298}
299
300/// A task to scan part of file.
301#[derive(Debug, Clone, Serialize, Deserialize, PartialEq, TypedBuilder)]
302#[builder(field_defaults(setter(prefix = "with_")))]
303pub struct FileScanTaskDeleteFile {
304    /// The delete file path
305    pub file_path: String,
306
307    /// The total size of the delete file in bytes, from the manifest entry.
308    pub file_size_in_bytes: u64,
309
310    /// delete file type
311    pub file_type: DataContentType,
312
313    /// The delete file's format, from the manifest entry. A `PositionDeletes` entry written as
314    /// `Puffin` is a V3 deletion vector; one written as `Parquet` is a position delete file.
315    pub file_format: DataFileFormat,
316
317    /// partition id
318    pub partition_spec_id: i32,
319
320    /// equality ids for equality deletes (null for anything other than equality-deletes)
321    #[builder(default)]
322    pub equality_ids: Option<Vec<i32>>,
323
324    /// For a deletion vector, the location of the data file whose rows it deletes. Required for
325    /// deletion vectors, and may also be set on a position delete file scoped to one data file.
326    #[serde(default)]
327    #[serde(skip_serializing_if = "Option::is_none")]
328    #[builder(default)]
329    pub referenced_data_file: Option<String>,
330
331    /// For a deletion vector, the offset of the blob within its Puffin file. Set only for
332    /// deletion vectors, where it locates the blob for direct access.
333    #[serde(default)]
334    #[serde(skip_serializing_if = "Option::is_none")]
335    #[builder(default)]
336    pub content_offset: Option<i64>,
337
338    /// For a deletion vector, the length in bytes of the blob within its Puffin file.
339    /// Required together with `content_offset`; both are absent for non-DV delete files.
340    #[serde(default)]
341    #[serde(skip_serializing_if = "Option::is_none")]
342    #[builder(default)]
343    pub content_size_in_bytes: Option<i64>,
344
345    /// The number of records in the delete file, from the manifest entry; for a deletion vector,
346    /// the cardinality of its bitmap. `None` only for a task not built from a manifest entry.
347    #[serde(default)]
348    #[serde(skip_serializing_if = "Option::is_none")]
349    #[builder(default)]
350    pub record_count: Option<u64>,
351
352    /// Key metadata for an encrypted delete file. When present, the reader uses this to
353    /// decrypt the file: for a Parquet equality or position delete file, this builds
354    /// `FileDecryptionProperties` (Parquet Modular Encryption); for a deletion vector, whose
355    /// Puffin file has no native encryption, this wraps the range read in an
356    /// `EncryptedInputFile` (AGS1 stream encryption).
357    ///
358    /// Same plaintext-DEK trust boundary as [`FileScanTask::key_metadata`]:
359    /// this is serialized into the scan plan and crosses the planner -> worker
360    /// channel in the clear for the standard encryption scheme.
361    #[serde(default)]
362    #[serde(skip_serializing_if = "Option::is_none")]
363    #[builder(default)]
364    pub key_metadata: Option<Box<[u8]>>,
365}
366
367mod _serde {
368    use std::sync::Arc;
369
370    use serde::{Deserialize, Serialize};
371
372    use super::{FileScanTask, FileScanTaskDeleteFile};
373    use crate::expr::BoundPredicate;
374    use crate::spec::{
375        DataFileFormat, Literal, NameMapping, PartitionSpec, RawLiteral, SchemaRef, StructType,
376        Type,
377    };
378    use crate::{Error, ErrorKind, Result};
379
380    #[derive(Deserialize)]
381    pub(super) struct FileScanTaskSerde {
382        file_size_in_bytes: u64,
383        start: u64,
384        length: u64,
385        record_count: Option<u64>,
386        first_row_id: Option<i64>,
387        data_sequence_number: Option<i64>,
388        data_file_path: String,
389        data_file_format: DataFileFormat,
390        schema: SchemaRef,
391        project_field_ids: Vec<i32>,
392        predicate: Option<BoundPredicate>,
393        deletes: Vec<FileScanTaskDeleteFile>,
394        #[serde(default)]
395        partition: Option<RawLiteral>,
396        #[serde(default)]
397        partition_spec: Option<Arc<PartitionSpec>>,
398        #[serde(default)]
399        name_mapping: Option<Arc<NameMapping>>,
400        #[serde(default)]
401        unified_partition_type: Option<Arc<StructType>>,
402        case_sensitive: bool,
403        #[serde(default)]
404        key_metadata: Option<Box<[u8]>>,
405    }
406
407    #[derive(Serialize)]
408    struct FileScanTaskRefSerde<'a> {
409        file_size_in_bytes: u64,
410        start: u64,
411        length: u64,
412        #[serde(skip_serializing_if = "Option::is_none")]
413        record_count: Option<u64>,
414        #[serde(skip_serializing_if = "Option::is_none")]
415        first_row_id: Option<i64>,
416        #[serde(skip_serializing_if = "Option::is_none")]
417        data_sequence_number: Option<i64>,
418        data_file_path: &'a str,
419        data_file_format: DataFileFormat,
420        schema: &'a SchemaRef,
421        project_field_ids: &'a [i32],
422        #[serde(skip_serializing_if = "Option::is_none")]
423        predicate: Option<&'a BoundPredicate>,
424        deletes: &'a [FileScanTaskDeleteFile],
425        #[serde(skip_serializing_if = "Option::is_none")]
426        partition: Option<RawLiteral>,
427        #[serde(skip_serializing_if = "Option::is_none")]
428        partition_spec: Option<&'a Arc<PartitionSpec>>,
429        #[serde(skip_serializing_if = "Option::is_none")]
430        name_mapping: Option<&'a Arc<NameMapping>>,
431        #[serde(skip_serializing_if = "Option::is_none")]
432        unified_partition_type: Option<&'a Arc<StructType>>,
433        case_sensitive: bool,
434        #[serde(skip_serializing_if = "Option::is_none")]
435        key_metadata: Option<&'a [u8]>,
436    }
437
438    fn partition_type(
439        partition_spec: Option<&PartitionSpec>,
440        schema: &crate::spec::Schema,
441    ) -> Result<Type> {
442        let partition_type = match partition_spec {
443            Some(partition_spec) => partition_spec.partition_type(schema)?,
444            None => PartitionSpec::unpartition_spec().partition_type(schema)?,
445        };
446        Ok(Type::Struct(partition_type))
447    }
448
449    impl<'a> TryFrom<&'a FileScanTask> for FileScanTaskRefSerde<'a> {
450        type Error = Error;
451
452        fn try_from(value: &'a FileScanTask) -> Result<Self> {
453            let partition = value
454                .partition
455                .as_ref()
456                .map(|partition| {
457                    let partition_type =
458                        partition_type(value.partition_spec.as_deref(), &value.schema)?;
459                    RawLiteral::try_from(Literal::Struct(partition.clone()), &partition_type)
460                })
461                .transpose()?;
462
463            Ok(Self {
464                file_size_in_bytes: value.file_size_in_bytes,
465                start: value.start,
466                length: value.length,
467                record_count: value.record_count,
468                first_row_id: value.first_row_id,
469                data_sequence_number: value.data_sequence_number,
470                data_file_path: &value.data_file_path,
471                data_file_format: value.data_file_format,
472                schema: &value.schema,
473                project_field_ids: &value.project_field_ids,
474                predicate: value.predicate.as_ref(),
475                deletes: &value.deletes,
476                partition,
477                partition_spec: value.partition_spec.as_ref(),
478                name_mapping: value.name_mapping.as_ref(),
479                unified_partition_type: value.unified_partition_type.as_ref(),
480                case_sensitive: value.case_sensitive,
481                key_metadata: value.key_metadata.as_deref(),
482            })
483        }
484    }
485
486    impl Serialize for FileScanTask {
487        fn serialize<S>(&self, serializer: S) -> std::result::Result<S::Ok, S::Error>
488        where S: serde::Serializer {
489            FileScanTaskRefSerde::try_from(self)
490                .map_err(serde::ser::Error::custom)?
491                .serialize(serializer)
492        }
493    }
494
495    impl TryFrom<FileScanTaskSerde> for FileScanTask {
496        type Error = Error;
497
498        fn try_from(value: FileScanTaskSerde) -> Result<Self> {
499            let partition = value
500                .partition
501                .map(|partition| {
502                    let partition_type =
503                        partition_type(value.partition_spec.as_deref(), &value.schema)?;
504                    match partition.try_into(&partition_type)? {
505                        Some(Literal::Struct(partition)) => Ok(partition),
506                        _ => Err(Error::new(
507                            ErrorKind::DataInvalid,
508                            "FileScanTask partition must be a struct",
509                        )),
510                    }
511                })
512                .transpose()?;
513
514            Self::builder()
515                .with_file_size_in_bytes(value.file_size_in_bytes)
516                .with_start(value.start)
517                .with_length(value.length)
518                .with_record_count(value.record_count)
519                .with_first_row_id(value.first_row_id)
520                .with_data_sequence_number(value.data_sequence_number)
521                .with_data_file_path(value.data_file_path)
522                .with_data_file_format(value.data_file_format)
523                .with_schema(value.schema)
524                .with_project_field_ids(value.project_field_ids)
525                .with_predicate(value.predicate)
526                .with_deletes(value.deletes)
527                .with_partition(partition)
528                .with_partition_spec(value.partition_spec)
529                .with_name_mapping(value.name_mapping)
530                .with_unified_partition_type(value.unified_partition_type)
531                .with_case_sensitive(value.case_sensitive)
532                .with_key_metadata(value.key_metadata)
533                .build()
534        }
535    }
536}
537
538#[cfg(test)]
539mod tests {
540    use super::*;
541    use crate::ErrorKind;
542    use crate::spec::{Literal, NestedField, PrimitiveType, Transform, Type};
543
544    fn build_file_scan_task(
545        schema: SchemaRef,
546        partition: Option<Struct>,
547        partition_spec: Option<Arc<PartitionSpec>>,
548    ) -> Result<FileScanTask> {
549        FileScanTask::builder()
550            .with_file_size_in_bytes(100)
551            .with_start(0)
552            .with_length(100)
553            .with_data_file_path("data_file_path".to_string())
554            .with_data_file_format(DataFileFormat::Parquet)
555            .with_schema(schema)
556            .with_project_field_ids(vec![])
557            .with_partition(partition)
558            .with_partition_spec(partition_spec)
559            .with_case_sensitive(false)
560            .build()
561    }
562
563    fn schema_and_spec(
564        primitive_type: PrimitiveType,
565        transform: Transform,
566    ) -> (SchemaRef, Arc<PartitionSpec>) {
567        let schema = Arc::new(
568            Schema::builder()
569                .with_fields(vec![Arc::new(NestedField::required(
570                    1,
571                    "x",
572                    Type::Primitive(primitive_type),
573                ))])
574                .build()
575                .unwrap(),
576        );
577        let partition_spec = Arc::new(
578            PartitionSpec::builder(schema.clone())
579                .add_partition_field("x", "x_partition", transform)
580                .unwrap()
581                .build()
582                .unwrap(),
583        );
584        (schema, partition_spec)
585    }
586
587    #[test]
588    fn test_file_scan_task_builder_rejects_non_empty_partition_without_spec() {
589        // Regression test for https://github.com/apache/iceberg-rust/issues/3130.
590        let err = build_file_scan_task(
591            Arc::new(Schema::builder().build().unwrap()),
592            Some(Struct::from_iter([Some(Literal::long(42))])),
593            None,
594        )
595        .unwrap_err();
596
597        assert_eq!(err.kind(), ErrorKind::DataInvalid);
598        assert_eq!(
599            err.message(),
600            "Non-empty FileScanTask partition requires a partition spec"
601        );
602    }
603
604    #[test]
605    fn test_file_scan_task_builder_accepts_empty_partition_without_spec() {
606        build_file_scan_task(
607            Arc::new(Schema::builder().build().unwrap()),
608            Some(Struct::empty()),
609            None,
610        )
611        .unwrap();
612    }
613
614    #[test]
615    fn test_file_scan_task_builder_rejects_partitioned_spec_without_partition() {
616        let (schema, partition_spec) = schema_and_spec(PrimitiveType::Long, Transform::Identity);
617
618        let err = build_file_scan_task(schema, None, Some(partition_spec)).unwrap_err();
619
620        assert_eq!(err.kind(), ErrorKind::DataInvalid);
621        assert_eq!(
622            err.message(),
623            "FileScanTask with a partitioned spec requires partition values"
624        );
625    }
626
627    #[test]
628    fn test_file_scan_task_builder_accepts_unpartitioned_spec_without_partition() {
629        build_file_scan_task(
630            Arc::new(Schema::builder().build().unwrap()),
631            None,
632            Some(Arc::new(PartitionSpec::unpartition_spec())),
633        )
634        .unwrap();
635    }
636
637    #[test]
638    fn test_file_scan_task_builder_rejects_partition_arity_mismatch() {
639        let (schema, partition_spec) = schema_and_spec(PrimitiveType::Long, Transform::Identity);
640
641        let err =
642            build_file_scan_task(schema, Some(Struct::empty()), Some(partition_spec)).unwrap_err();
643
644        assert_eq!(err.kind(), ErrorKind::DataInvalid);
645        assert!(err.message().contains("partition has 0 fields"));
646        assert!(err.message().contains("partition spec has 1 fields"));
647    }
648
649    #[test]
650    fn test_file_scan_task_builder_rejects_dropped_partition_source_column() {
651        let (_historical_schema, partition_spec) =
652            schema_and_spec(PrimitiveType::Long, Transform::Identity);
653        let current_schema = Arc::new(
654            Schema::builder()
655                .with_fields(vec![Arc::new(NestedField::required(
656                    2,
657                    "y",
658                    Type::Primitive(PrimitiveType::String),
659                ))])
660                .build()
661                .unwrap(),
662        );
663
664        let err = build_file_scan_task(
665            current_schema,
666            Some(Struct::from_iter([Some(Literal::long(42))])),
667            Some(partition_spec),
668        )
669        .unwrap_err();
670
671        assert_eq!(err.kind(), ErrorKind::Unexpected);
672        assert!(err.message().contains("No column with source column id 1"));
673    }
674
675    #[test]
676    fn test_file_scan_task_builder_rejects_partition_spec_incompatible_with_schema() {
677        let (_historical_schema, partition_spec) =
678            schema_and_spec(PrimitiveType::Timestamp, Transform::Day);
679        let current_schema = Arc::new(
680            Schema::builder()
681                .with_fields(vec![Arc::new(NestedField::required(
682                    1,
683                    "x",
684                    Type::Primitive(PrimitiveType::String),
685                ))])
686                .build()
687                .unwrap(),
688        );
689
690        let err = build_file_scan_task(
691            current_schema,
692            Some(Struct::from_iter([Some(Literal::date(20_000))])),
693            Some(partition_spec),
694        )
695        .unwrap_err();
696
697        assert_eq!(err.kind(), ErrorKind::DataInvalid);
698    }
699}
iceberg/scan/task.rs

iceberg/scan/
task.rs
