fix: Move partition metadata types to data_types2

Cargo.lock

@@ -1174,6 +1174,7 @@ version = "0.1.0"
 dependencies = [
  "data_types",
  "influxdb_line_protocol",
+ "observability_deps",
  "ordered-float 3.0.0",
  "percent-encoding",
  "schema",
@@ -4522,6 +4523,7 @@ dependencies = [
  "criterion",
  "croaring",
  "data_types",
+ "data_types2",
  "datafusion 0.1.0",
  "either",
  "hashbrown 0.12.0",

data_types2/Cargo.toml

@@ -7,6 +7,7 @@ description = "Shared data types in the IOx NG architecture"
 [dependencies]
 data_types = { path = "../data_types" }
 influxdb_line_protocol = { path = "../influxdb_line_protocol" }
+observability_deps = { path = "../observability_deps" }
 ordered-float = "3"
 percent-encoding = "2.1.0"
 schema = { path = "../schema" }

dml/src/lib.rs

@@ -12,11 +12,10 @@
 )]
 
 use data_types::{
-    partition_metadata::{StatValues, Statistics},
     router::{ShardConfig, ShardId},
     sequence::Sequence,
 };
-use data_types2::{DeletePredicate, NonEmptyString};
+use data_types2::{DeletePredicate, NonEmptyString, StatValues, Statistics};
 use hashbrown::HashMap;
 use iox_time::Time;
 use mutable_batch::MutableBatch;

mutable_batch/src/column.rs

@@ -1,23 +1,18 @@
 //! A [`Column`] stores the rows for a given column name
 
-use std::fmt::Formatter;
-use std::mem;
-use std::sync::Arc;
-
-use arrow::error::ArrowError;
 use arrow::{
     array::{
         ArrayDataBuilder, ArrayRef, BooleanArray, Float64Array, Int64Array,
         TimestampNanosecondArray, UInt64Array,
     },
     datatypes::DataType,
+    error::ArrowError,
 };
-use snafu::{ResultExt, Snafu};
-
-use arrow_util::bitset::BitSet;
-use arrow_util::string::PackedStringArray;
-use data_types::partition_metadata::{StatValues, Statistics};
+use arrow_util::{bitset::BitSet, string::PackedStringArray};
+use data_types2::{StatValues, Statistics};
 use schema::{InfluxColumnType, InfluxFieldType, TIME_DATA_TYPE};
+use snafu::{ResultExt, Snafu};
+use std::{fmt::Formatter, mem, sync::Arc};
 
 /// A "dictionary ID" (DID) is a compact numeric representation of an interned
 /// string in the dictionary. The same string always maps the same DID.

mutable_batch/src/writer.rs

@@ -3,7 +3,7 @@
 use crate::column::{Column, ColumnData, INVALID_DID};
 use crate::MutableBatch;
 use arrow_util::bitset::{iter_set_positions, iter_set_positions_with_offset, BitSet};
-use data_types::partition_metadata::{IsNan, StatValues, Statistics};
+use data_types2::{IsNan, StatValues, Statistics};
 use schema::{InfluxColumnType, InfluxFieldType};
 use snafu::Snafu;
 use std::num::NonZeroU64;

mutable_batch/tests/extend.rs

@@ -1,10 +1,8 @@
 use arrow_util::assert_batches_eq;
-use data_types::partition_metadata::{StatValues, Statistics};
-use mutable_batch::writer::Writer;
-use mutable_batch::MutableBatch;
+use data_types2::{StatValues, Statistics};
+use mutable_batch::{writer::Writer, MutableBatch};
 use schema::selection::Selection;
-use std::collections::BTreeMap;
-use std::num::NonZeroU64;
+use std::{collections::BTreeMap, num::NonZeroU64};
 
 #[test]
 fn test_extend() {

mutable_batch/tests/extend_range.rs

@@ -1,10 +1,8 @@
 use arrow_util::assert_batches_eq;
-use data_types::partition_metadata::{StatValues, Statistics};
-use mutable_batch::writer::Writer;
-use mutable_batch::MutableBatch;
+use data_types2::{StatValues, Statistics};
+use mutable_batch::{writer::Writer, MutableBatch};
 use schema::selection::Selection;
-use std::collections::BTreeMap;
-use std::num::NonZeroU64;
+use std::{collections::BTreeMap, num::NonZeroU64};
 
 #[test]
 fn test_extend_range() {

mutable_batch/tests/writer.rs

@@ -1,6 +1,6 @@
 use arrow_util::assert_batches_eq;
-use data_types::partition_metadata::{StatValues, Statistics};
 use data_types::write_summary::TimestampSummary;
+use data_types2::{StatValues, Statistics};
 use mutable_batch::writer::Writer;
 use mutable_batch::MutableBatch;
 use schema::selection::Selection;

mutable_batch/tests/writer_fuzz.rs

@@ -1,3 +1,16 @@
+use arrow::{
+    array::{
+        ArrayRef, BooleanArray, Float64Array, Int64Array, StringArray, TimestampNanosecondArray,
+        UInt64Array,
+    },
+    record_batch::RecordBatch,
+};
+use arrow_util::bitset::BitSet;
+use data_types2::{IsNan, PartitionTemplate, StatValues, Statistics, TemplatePart};
+use hashbrown::HashSet;
+use mutable_batch::{writer::Writer, MutableBatch, PartitionWrite, WritePayload};
+use rand::prelude::*;
+use schema::selection::Selection;
 /// A fuzz test of the [`mutable_batch::Writer`] interface:
 ///
 /// - column writes - `write_i64`, `write_tag`, etc...
@@ -5,25 +18,7 @@
 /// - batch writes with ranges - `write_batch_ranges`
 ///
 /// Verifies that the rows and statistics are as expected after a number of interleaved writes
-use std::collections::BTreeMap;
-use std::num::NonZeroU64;
-use std::ops::Range;
-use std::sync::Arc;
-
-use arrow::array::{
-    ArrayRef, BooleanArray, Float64Array, Int64Array, StringArray, TimestampNanosecondArray,
-    UInt64Array,
-};
-use arrow::record_batch::RecordBatch;
-use hashbrown::HashSet;
-use rand::prelude::*;
-
-use arrow_util::bitset::BitSet;
-use data_types::partition_metadata::{IsNan, StatValues, Statistics};
-use data_types2::{PartitionTemplate, TemplatePart};
-use mutable_batch::writer::Writer;
-use mutable_batch::{MutableBatch, PartitionWrite, WritePayload};
-use schema::selection::Selection;
+use std::{collections::BTreeMap, num::NonZeroU64, ops::Range, sync::Arc};
 
 fn make_rng() -> StdRng {
     let seed = rand::rngs::OsRng::default().next_u64();

parquet_file/src/chunk.rs

@@ -3,17 +3,14 @@ use crate::{
     storage::Storage,
     ParquetFilePath,
 };
-use data_types::{
-    partition_metadata::{Statistics, TableSummary},
-    timestamp::{TimestampMinMax, TimestampRange},
+use data_types2::{
+    ParquetFile, ParquetFileWithMetadata, Statistics, TableSummary, TimestampMinMax, TimestampRange,
 };
-use data_types2::{ParquetFile, ParquetFileWithMetadata};
 use datafusion::physical_plan::SendableRecordBatchStream;
 use object_store::DynObjectStore;
 use observability_deps::tracing::*;
 use predicate::Predicate;
-use schema::selection::Selection;
-use schema::{Schema, TIME_COLUMN_NAME};
+use schema::{selection::Selection, Schema, TIME_COLUMN_NAME};
 use snafu::{ResultExt, Snafu};
 use std::{collections::BTreeSet, mem, sync::Arc};
 

parquet_file/src/metadata.rs

@@ -86,9 +86,9 @@
 //! [Apache Parquet]: https://parquet.apache.org/
 //! [Apache Thrift]: https://thrift.apache.org/
 //! [Thrift Compact Protocol]: https://github.com/apache/thrift/blob/master/doc/specs/thrift-compact-protocol.md
-use data_types::partition_metadata::{ColumnSummary, InfluxDbType, StatValues, Statistics};
 use data_types2::{
-    NamespaceId, ParquetFileParams, PartitionId, SequenceNumber, SequencerId, TableId, Timestamp,
+    ColumnSummary, InfluxDbType, NamespaceId, ParquetFileParams, PartitionId, SequenceNumber,
+    SequencerId, StatValues, Statistics, TableId, Timestamp,
 };
 use generated_types::influxdata::iox::ingester::v1 as proto;
 use iox_time::Time;
@@ -106,8 +106,10 @@ use parquet::{
     schema::types::SchemaDescriptor as ParquetSchemaDescriptor,
 };
 use prost::Message;
-use schema::sort::{SortKey, SortKeyBuilder};
-use schema::{InfluxColumnType, InfluxFieldType, Schema};
+use schema::{
+    sort::{SortKey, SortKeyBuilder},
+    InfluxColumnType, InfluxFieldType, Schema,
+};
 use snafu::{ensure, OptionExt, ResultExt, Snafu};
 use std::{convert::TryInto, sync::Arc};
 use thrift::protocol::{TCompactInputProtocol, TCompactOutputProtocol, TOutputProtocol};

querier/src/chunk/query_access.rs

@@ -1,7 +1,8 @@
 use std::sync::Arc;
 
-use data_types::timestamp::TimestampMinMax;
-use data_types2::{ChunkAddr, ChunkId, ChunkOrder, DeletePredicate, PartitionId, TableSummary};
+use data_types2::{
+    ChunkAddr, ChunkId, ChunkOrder, DeletePredicate, PartitionId, TableSummary, TimestampMinMax,
+};
 use observability_deps::tracing::debug;
 use predicate::PredicateMatch;
 use query::{QueryChunk, QueryChunkError, QueryChunkMeta};

query/src/provider/overlap.rs

@@ -1,21 +1,18 @@
-//! Contains the algorithm to determine which chunks may contain
-//! "duplicate" primary keys (that is where data with the same
-//! combination of "tag" columns and timestamp in the InfluxDB
-//! DataModel have been written in via multiple distinct line protocol
-//! writes (and thus are stored in separate rows)
+//! Contains the algorithm to determine which chunks may contain "duplicate" primary keys (that is
+//! where data with the same combination of "tag" columns and timestamp in the InfluxDB DataModel
+//! have been written in via multiple distinct line protocol writes (and thus are stored in
+//! separate rows)
 
-use data_types::{
-    partition_metadata::{ColumnSummary, StatOverlap, Statistics},
-    timestamp::TimestampMinMax,
+use crate::{QueryChunk, QueryChunkMeta};
+use data_types2::{
+    ColumnSummary, DeletePredicate, ParquetFileWithMetadata, PartitionId, StatOverlap, Statistics,
+    TableSummary, TimestampMinMax,
 };
-use data_types2::{DeletePredicate, ParquetFileWithMetadata, PartitionId, TableSummary};
 use observability_deps::tracing::debug;
 use schema::{sort::SortKey, Schema, TIME_COLUMN_NAME};
 use snafu::Snafu;
 use std::{cmp::Ordering, sync::Arc};
 
-use crate::{QueryChunk, QueryChunkMeta};
-
 #[derive(Debug, Snafu)]
 pub enum Error {
     #[snafu(display(

query/src/provider/physical.rs

@@ -1,9 +1,9 @@
 //! Implementation of a DataFusion PhysicalPlan node across partition chunks
 
-use std::{fmt, sync::Arc};
-
+use super::adapter::SchemaAdapterStream;
+use crate::{exec::IOxSessionContext, QueryChunk};
 use arrow::datatypes::SchemaRef;
-use data_types::partition_metadata::TableSummary;
+use data_types2::TableSummary;
 use datafusion::{
     error::DataFusionError,
     execution::context::TaskContext,
@@ -14,13 +14,9 @@ use datafusion::{
     },
 };
 use observability_deps::tracing::debug;
-use schema::selection::Selection;
-use schema::Schema;
-
-use crate::{exec::IOxSessionContext, QueryChunk};
 use predicate::Predicate;
-
-use super::adapter::SchemaAdapterStream;
+use schema::{selection::Selection, Schema};
+use std::{fmt, sync::Arc};
 
 /// Implements the DataFusion physical plan interface
 #[derive(Debug)]

query/src/pruning.rs

@@ -1,23 +1,22 @@
 //! Implementation of statistics based pruning
 
-use std::sync::Arc;
-
-use arrow::array::{
-    ArrayRef, BooleanArray, DictionaryArray, Float64Array, Int64Array, StringArray, UInt64Array,
+use crate::QueryChunk;
+use arrow::{
+    array::{
+        ArrayRef, BooleanArray, DictionaryArray, Float64Array, Int64Array, StringArray, UInt64Array,
+    },
+    datatypes::{DataType, Int32Type, TimeUnit},
 };
-use arrow::datatypes::{DataType, Int32Type, TimeUnit};
-
-use data_types::partition_metadata::{StatValues, Statistics};
+use data_types2::{StatValues, Statistics};
 use datafusion::{
     logical_plan::Column,
     physical_optimizer::pruning::{PruningPredicate, PruningStatistics},
 };
 use observability_deps::tracing::{debug, trace};
 use predicate::Predicate;
-use schema::Schema;
-
-use crate::QueryChunk;
 use query_functions::group_by::Aggregate;
+use schema::Schema;
+use std::sync::Arc;
 
 /// Something that cares to be notified when pruning of chunks occurs
 pub trait PruningObserver {

query/src/statistics.rs

@@ -1,8 +1,6 @@
 //! Code to translate IOx statistics to DataFusion statistics
 
-use data_types::partition_metadata::{
-    ColumnSummary, InfluxDbType, Statistics as IOxStatistics, TableSummary,
-};
+use data_types2::{ColumnSummary, InfluxDbType, Statistics as IOxStatistics, TableSummary};
 use datafusion::{
     physical_plan::{ColumnStatistics, Statistics as DFStatistics},
     scalar::ScalarValue,
@@ -107,11 +105,10 @@ fn df_from_iox_col(col: &ColumnSummary) -> ColumnStatistics {
 
 #[cfg(test)]
 mod test {
-    use std::num::NonZeroU64;
-
     use super::*;
-    use data_types::partition_metadata::{InfluxDbType, StatValues};
+    use data_types2::{InfluxDbType, StatValues};
     use schema::{builder::SchemaBuilder, InfluxFieldType};
+    use std::num::NonZeroU64;
 
     macro_rules! assert_nice_eq {
         ($actual:ident, $expected:ident) => {

query/src/test.rs

@@ -283,7 +283,7 @@ impl TestChunk {
         Self {
             table_name,
             schema: Arc::new(SchemaBuilder::new().build().unwrap()),
-            table_summary: TableSummary::new(),
+            table_summary: TableSummary::default(),
             id: ChunkId::new_test(0),
             may_contain_pk_duplicates: Default::default(),
             predicates: Default::default(),

read_buffer/Cargo.toml

@@ -15,6 +15,7 @@ arrow = { version = "13", features = ["prettyprint"] }
 arrow_util = { path = "../arrow_util" }
 croaring = "0.6"
 data_types = { path = "../data_types" }
+data_types2 = { path = "../data_types2" }
 datafusion = { path = "../datafusion" }
 either = "1.6.1"
 hashbrown = "0.12"

read_buffer/src/chunk.rs

@@ -5,11 +5,10 @@ use crate::{
     table::{self, Table},
 };
 use arrow::{error::ArrowError, record_batch::RecordBatch};
-use data_types::{chunk_metadata::ChunkColumnSummary, partition_metadata::TableSummary};
-
+use data_types::chunk_metadata::ChunkColumnSummary;
+use data_types2::TableSummary;
 use observability_deps::tracing::debug;
-use schema::selection::Selection;
-use schema::{builder::Error as SchemaError, Schema};
+use schema::{builder::Error as SchemaError, selection::Selection, Schema};
 use snafu::{ResultExt, Snafu};
 use std::{
     collections::{BTreeMap, BTreeSet},
@@ -527,11 +526,10 @@ mod test {
             Int32Type,
         },
     };
-    use data_types::partition_metadata::{ColumnSummary, InfluxDbType, StatValues, Statistics};
+    use data_types2::{ColumnSummary, InfluxDbType, StatValues, Statistics};
     use metric::{Attributes, MetricKind, Observation, ObservationSet, RawReporter};
     use schema::builder::SchemaBuilder;
-    use std::iter::FromIterator;
-    use std::{num::NonZeroU64, sync::Arc};
+    use std::{iter::FromIterator, num::NonZeroU64, sync::Arc};
 
     // helper to make the `add_remove_tables` test simpler to read.
     fn gen_recordbatch() -> RecordBatch {

read_buffer/src/schema.rs

@@ -241,8 +241,8 @@ impl ColumnType {
         }
     }
 
-    pub fn as_influxdb_type(&self) -> Option<data_types::partition_metadata::InfluxDbType> {
-        use data_types::partition_metadata::InfluxDbType;
+    pub fn as_influxdb_type(&self) -> Option<data_types2::InfluxDbType> {
+        use data_types2::InfluxDbType;
         match self {
             Self::Tag(_) => Some(InfluxDbType::Tag),
             Self::Field(_) => Some(InfluxDbType::Field),

read_buffer/src/table.rs

@@ -6,7 +6,8 @@ use crate::{
     BinaryExpr,
 };
 use arrow::record_batch::RecordBatch;
-use data_types::{chunk_metadata::ChunkColumnSummary, partition_metadata::TableSummary};
+use data_types::chunk_metadata::ChunkColumnSummary;
+use data_types2::TableSummary;
 use parking_lot::RwLock;
 use schema::selection::Selection;
 use snafu::{ensure, Snafu};
@@ -782,7 +783,7 @@ impl MetaData {
     }
 
     pub fn to_summary(&self) -> TableSummary {
-        use data_types::partition_metadata::{ColumnSummary, StatValues, Statistics};
+        use data_types2::{ColumnSummary, StatValues, Statistics};
         let columns = self
             .columns
             .iter()
@@ -850,8 +851,8 @@ impl MetaData {
 fn make_null_stats(
     total_count: u64,
     logical_data_type: &LogicalDataType,
-) -> data_types::partition_metadata::Statistics {
-    use data_types::partition_metadata::{StatValues, Statistics};
+) -> data_types2::Statistics {
+    use data_types2::{StatValues, Statistics};
     use LogicalDataType::*;
 
     match logical_data_type {
@@ -1103,9 +1104,6 @@ impl std::fmt::Display for DisplayReadAggregateResults<'_> {
 
 #[cfg(test)]
 mod test {
-    use arrow::array::BooleanArray;
-    use data_types::partition_metadata::{StatValues, Statistics};
-
     use super::*;
     use crate::{
         column::Column,
@@ -1113,6 +1111,8 @@ mod test {
         schema::{self, LogicalDataType},
         value::{AggregateVec, OwnedValue, Scalar},
     };
+    use arrow::array::BooleanArray;
+    use data_types2::{StatValues, Statistics};
 
     #[test]
     fn meta_data_update_with() {