From 405a6a691b2edb5e6e6a44c0b690c8e70cd43d69 Mon Sep 17 00:00:00 2001
From: Nga Tran <ntran@influxdata.com>
Date: Fri, 2 Jul 2021 17:57:48 -0400
Subject: [PATCH] feat: intial implementation of #1886: avoid resort if
 appropriate

---
 Cargo.lock                       |  1 +
 data_types/Cargo.toml            |  1 +
 data_types/src/chunk_metadata.rs | 41 +++++++++++++++++++++++++++++---
 server/src/db/chunk.rs           |  6 ++++-
 4 files changed, 45 insertions(+), 4 deletions(-)

diff --git a/Cargo.lock b/Cargo.lock
index 65d69cab8e..3f58c83b62 100644
--- a/Cargo.lock
+++ b/Cargo.lock
@@ -818,6 +818,7 @@ name = "data_types"
 version = "0.1.0"
 dependencies = [
  "chrono",
+ "indexmap",
  "influxdb_line_protocol",
  "observability_deps",
  "percent-encoding",
diff --git a/data_types/Cargo.toml b/data_types/Cargo.toml
index bfd1331a30..1bb7d0d88e 100644
--- a/data_types/Cargo.toml
+++ b/data_types/Cargo.toml
@@ -8,6 +8,7 @@ readme = "README.md"
 
 [dependencies] # In alphabetical order
 chrono = { version = "0.4", features = ["serde"] }
+indexmap = "1.6"
 influxdb_line_protocol = { path = "../influxdb_line_protocol" }
 percent-encoding = "2.1.0"
 regex = "1.4"
diff --git a/data_types/src/chunk_metadata.rs b/data_types/src/chunk_metadata.rs
index 63e2b0bb82..e713c372c0 100644
--- a/data_types/src/chunk_metadata.rs
+++ b/data_types/src/chunk_metadata.rs
@@ -1,5 +1,7 @@
 //! Module contains a representation of chunk metadata
-use std::sync::Arc;
+use std::{fmt, sync::Arc};
+
+use indexmap::IndexMap;
 
 use chrono::{DateTime, Utc};
 use serde::{Deserialize, Serialize};
@@ -18,14 +20,47 @@ pub struct ChunkAddr {
 
     /// The ID of the chunk
     pub chunk_id: u32,
+
+    // Sort key of this chunk
+    pub sort_key: Arc<SortKey>,
+}
+
+/// Temporary - https://github.com/apache/arrow-rs/pull/425
+#[derive(Debug, Clone, Copy, Eq, PartialEq, Serialize, Deserialize)]
+pub struct SortOptions {
+    /// Whether to sort in descending order
+    pub descending: bool,
+    /// Whether to sort nulls first
+    pub nulls_first: bool,
+}
+impl fmt::Display for SortOptions {
+    fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
+        write!(f, "Descending: {}, NUll first: {}", self.descending, self.nulls_first)
+    }
+}
+
+#[derive(Debug, Clone, Eq, PartialEq, Serialize, Deserialize)]
+pub struct SortKey {
+    columns: Vec<(String, SortOptions)>,
+}
+
+impl fmt::Display for SortKey {
+    fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
+        let mut sort_key = "Sort Key:\n".to_string();
+        for col in &self.columns {
+            let s = format!("{}, {}", col.0, col.1);
+            sort_key = sort_key + &s;
+        }
+        write!(f, "{}", sort_key)
+    }
 }
 
 impl std::fmt::Display for ChunkAddr {
     fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
         write!(
             f,
-            "Chunk('{}':'{}':'{}':{})",
-            self.db_name, self.table_name, self.partition_key, self.chunk_id
+            "Chunk('{}':'{}':'{}':{}:'{}')",
+            self.db_name, self.table_name, self.partition_key, self.chunk_id, self.sort_key
         )
     }
 }
diff --git a/server/src/db/chunk.rs b/server/src/db/chunk.rs
index 012417ba40..88ebdf0376 100644
--- a/server/src/db/chunk.rs
+++ b/server/src/db/chunk.rs
@@ -432,7 +432,11 @@ impl QueryChunk for DbChunk {
         }
     }
 
-    // TODOs: return the right value. For now the chunk is assumed to be not sorted
+    /// Returns true if the chunk is sorted on its pk
+    /// Since data is compacted prior being moved to RUBs, data in RUBs and OBs
+    /// should be sorted on their PK as the results of compacting.
+    /// However, since we current sorted data based on their cardinality (see compute_sort_key),
+    /// 2 different chunks may be sorted on different order of key columns.
     fn is_sorted_on_pk(&self) -> bool {
         match &self.state {
             State::MutableBuffer { .. } => false,