Update the docs for EARLIEST_BY/LATEST_BY aggregators with the newly added numeric capabilities

docs/multi-stage-query/concepts.md

@@ -200,8 +200,8 @@ rollup-related metadata into the generated segments. Other applications can then
 queries](../querying/segmentmetadataquery.md) to retrieve rollup-related information.
 
 The following [aggregation functions](../querying/sql-aggregations.md) are supported for rollup at ingestion time:
-`COUNT` (but switch to `SUM` at query time), `SUM`, `MIN`, `MAX`, `EARLIEST` and `EARLIEST_BY` ([string only](known-issues.md#select-statement)),
-`LATEST` and `LATEST_BY` ([string only](known-issues.md#select-statement)), `APPROX_COUNT_DISTINCT`, `APPROX_COUNT_DISTINCT_BUILTIN`,
+`COUNT` (but switch to `SUM` at query time), `SUM`, `MIN`, `MAX`, `EARLIEST` and `EARLIEST_BY`,
+`LATEST` and `LATEST_BY`, `APPROX_COUNT_DISTINCT`, `APPROX_COUNT_DISTINCT_BUILTIN`,
 `APPROX_COUNT_DISTINCT_DS_HLL`, `APPROX_COUNT_DISTINCT_DS_THETA`, and `DS_QUANTILES_SKETCH` (but switch to
 `APPROX_QUANTILE_DS` at query time). Do not use `AVG`; instead, use `SUM` and `COUNT` at ingest time and compute the
 quotient at query time.

docs/multi-stage-query/known-issues.md

@@ -42,11 +42,6 @@ an [UnknownError](./reference.md#error_UnknownError) with a message including "N
 - `GROUPING SETS` are not implemented. Queries using these features return a
   [QueryNotSupported](reference.md#error_QueryNotSupported) error.
 
-- The numeric varieties of the `EARLIEST` and `LATEST` aggregators do not work properly. Attempting to use the numeric
-  varieties of these aggregators lead to an error like
-  `java.lang.ClassCastException: class java.lang.Double cannot be cast to class org.apache.druid.collections.SerializablePair`.
-  The string varieties, however, do work properly.
-
 ## `INSERT` and `REPLACE` Statements
 
 - The `INSERT` and `REPLACE` statements with column lists, like `INSERT INTO tbl (a, b, c) SELECT ...`, is not implemented.

docs/querying/aggregations.md

@@ -177,10 +177,9 @@ Example:
 
 The first and last aggregators determine the metric values that respectively correspond to the earliest and latest values of a time column.
 
-Do not use first and last aggregators for the double, float, and long types in an ingestion spec. They are only supported for queries.
-The string-typed aggregators, `stringFirst` and `stringLast`, are supported for both ingestion and querying.
-
-Queries with first or last aggregators on a segment created with rollup return the rolled up value, not the first or last value from the raw ingested data.
+Queries with first or last aggregators on a segment created with rollup return the rolled up value, not the first or last value from the 
+raw ingested data. The `timeColumn` will get ignored in such cases, and the aggregation will use the original value of the time column
+stored at the time the segment was created.
 
 #### Numeric first and last aggregators
 

docs/querying/sql-aggregations.md

@@ -87,9 +87,9 @@ In the aggregation functions supported by Druid, only `COUNT`, `ARRAY_AGG`, and
 |`STDDEV_SAMP(expr)`|Computes standard deviation sample of `expr`. See [stats extension](../development/extensions-core/stats.md) documentation for additional details.|`null` or `0` if `druid.generic.useDefaultValueForNull=true` (legacy mode)|
 |`STDDEV(expr)`|Computes standard deviation sample of `expr`. See [stats extension](../development/extensions-core/stats.md) documentation for additional details.|`null` or `0` if `druid.generic.useDefaultValueForNull=true` (legacy mode)|
 |`EARLIEST(expr, [maxBytesPerValue])`|Returns the earliest value of `expr`.<br />If `expr` comes from a relation with a timestamp column (like `__time` in a Druid datasource), the "earliest" is taken from the row with the overall earliest non-null value of the timestamp column.<br />If the earliest non-null value of the timestamp column appears in multiple rows, the `expr` may be taken from any of those rows. If `expr` does not come from a relation with a timestamp, then it is simply the first value encountered.<br /><br />If `expr` is a string or complex type `maxBytesPerValue` amount of space is allocated for the aggregation. Strings longer than this limit are truncated. The  `maxBytesPerValue` parameter should be set as low as possible, since high values will lead to wasted memory.<br/>If `maxBytesPerValue`is omitted; it defaults to `1024`. |`null` or `0`/`''` if `druid.generic.useDefaultValueForNull=true` (legacy mode)|
-|`EARLIEST_BY(expr, timestampExpr, [maxBytesPerValue])`|Returns the earliest value of `expr`.<br />The earliest value of `expr` is taken from the row with the overall earliest non-null value of `timestampExpr`. <br />If the earliest non-null value of `timestampExpr` appears in multiple rows, the `expr` may be taken from any of those rows.<br /><br />If `expr` is a string or complex type `maxBytesPerValue` amount of space is allocated for the aggregation. Strings longer than this limit are truncated. The  `maxBytesPerValue` parameter should be set as low as possible, since high values will lead to wasted memory.<br/>If `maxBytesPerValue`is omitted; it defaults to `1024`. |`null` or `0`/`''` if `druid.generic.useDefaultValueForNull=true` (legacy mode)|
+|`EARLIEST_BY(expr, timestampExpr, [maxBytesPerValue])`|Returns the earliest value of `expr`.<br />The earliest value of `expr` is taken from the row with the overall earliest non-null value of `timestampExpr`. <br />If the earliest non-null value of `timestampExpr` appears in multiple rows, the `expr` may be taken from any of those rows.<br /><br />If `expr` is a string or complex type `maxBytesPerValue` amount of space is allocated for the aggregation. Strings longer than this limit are truncated. The  `maxBytesPerValue` parameter should be set as low as possible, since high values will lead to wasted memory.<br/>If `maxBytesPerValue`is omitted; it defaults to `1024`.<br /><br />Use `EARLIEST` instead of `EARLIEST_BY` on a table that has rollup enabled and was created with any variant of `EARLIEST`, `LATEST`, `EARLIEST_BY`, or `LATEST_BY`. In these cases, the intermediate type already stores the timestamp, and Druid ignores the value passed in `timestampExpr`. |`null` or `0`/`''` if `druid.generic.useDefaultValueForNull=true` (legacy mode)|
 |`LATEST(expr, [maxBytesPerValue])`|Returns the latest value of `expr`<br />The `expr` must come from a relation with a timestamp column (like `__time` in a Druid datasource) and the "latest" is taken from the row with the overall latest non-null value of the timestamp column.<br />If the latest non-null value of the timestamp column appears in multiple rows, the `expr` may be taken from any of those rows.<br /><br />If `expr` is a string or complex type `maxBytesPerValue` amount of space is allocated for the aggregation. Strings longer than this limit are truncated. The  `maxBytesPerValue` parameter should be set as low as possible, since high values will lead to wasted memory.<br/>If `maxBytesPerValue`is omitted; it defaults to `1024`. |`null` or `0`/`''` if `druid.generic.useDefaultValueForNull=true` (legacy mode)|
-|`LATEST_BY(expr, timestampExpr, [maxBytesPerValue])`|Returns the latest value of `expr`.<br />The latest value of `expr` is taken from the row with the overall latest non-null value of `timestampExpr`.<br />If the overall latest non-null value of `timestampExpr` appears in multiple rows, the `expr` may be taken from any of those rows.<br /><br />If `expr` is a string or complex type `maxBytesPerValue` amount of space is allocated for the aggregation. Strings longer than this limit are truncated. The `maxBytesPerValue` parameter should be set as low as possible, since high values will lead to wasted memory.<br/>If `maxBytesPerValue`is omitted; it defaults to `1024`. |`null` or `0`/`''` if `druid.generic.useDefaultValueForNull=true` (legacy mode)|
+|`LATEST_BY(expr, timestampExpr, [maxBytesPerValue])`|Returns the latest value of `expr`.<br />The latest value of `expr` is taken from the row with the overall latest non-null value of `timestampExpr`.<br />If the overall latest non-null value of `timestampExpr` appears in multiple rows, the `expr` may be taken from any of those rows.<br /><br />If `expr` is a string or complex type `maxBytesPerValue` amount of space is allocated for the aggregation. Strings longer than this limit are truncated. The `maxBytesPerValue` parameter should be set as low as possible, since high values will lead to wasted memory.<br/>If `maxBytesPerValue`is omitted; it defaults to `1024`.<br /><br />Use `LATEST` instead of `LATEST_BY` on a table that has rollup enabled and was created with any variant of `EARLIEST`, `LATEST`, `EARLIEST_BY`, or `LATEST_BY`. In these cases, the intermediate type already stores the timestamp, and Druid ignores the value passed in `timestampExpr`. |`null` or `0`/`''` if `druid.generic.useDefaultValueForNull=true` (legacy mode)|
 |`ANY_VALUE(expr, [maxBytesPerValue, [aggregateMultipleValues]])`|Returns any value of `expr` including null. This aggregator can simplify and optimize the performance by returning the first encountered value (including `null`).<br /><br />If `expr` is a string or complex type `maxBytesPerValue` amount of space is allocated for the aggregation. Strings longer than this limit are truncated. The `maxBytesPerValue` parameter should be set as low as possible, since high values will lead to wasted memory.<br/>If `maxBytesPerValue` is omitted; it defaults to `1024`. `aggregateMultipleValues` is an optional boolean flag controls the behavior of aggregating a [multi-value dimension](./multi-value-dimensions.md). `aggregateMultipleValues` is set as true by default and returns the stringified array in case of a multi-value dimension. By setting it to false, function will return first value instead. |`null` or `0`/`''` if `druid.generic.useDefaultValueForNull=true` (legacy mode)|
 |`GROUPING(expr, expr...)`|Returns a number to indicate which groupBy dimension is included in a row, when using `GROUPING SETS`. Refer to [additional documentation](aggregations.md#grouping-aggregator) on how to infer this number.|N/A|
 |`ARRAY_AGG(expr, [size])`|Collects all values of `expr` into an ARRAY, including null values, with `size` in bytes limit on aggregation size (default of 1024 bytes). If the aggregated array grows larger than the maximum size in bytes, the query will fail. Use of `ORDER BY` within the `ARRAY_AGG` expression is not currently supported, and the ordering of results within the output array may vary depending on processing order.|`null`|

sql/src/main/java/org/apache/druid/sql/calcite/aggregation/builtin/EarliestLatestAnySqlAggregator.java

@@ -229,7 +229,7 @@ public Aggregation toDruidAggregation(
                           );
     }
 
-    final String fieldName = getColumnName(plannerContext, virtualColumnRegistry, args.get(0), rexNodes.get(0));
+    final String fieldName = getColumnName(virtualColumnRegistry, args.get(0), rexNodes.get(0));
 
     if (!inputAccessor.getInputRowSignature().contains(ColumnHolder.TIME_COLUMN_NAME)
         && (aggregatorType == AggregatorType.LATEST || aggregatorType == AggregatorType.EARLIEST)) {
@@ -291,7 +291,6 @@ public Aggregation toDruidAggregation(
   }
 
   static String getColumnName(
-      PlannerContext plannerContext,
       VirtualColumnRegistry virtualColumnRegistry,
       DruidExpression arg,
       RexNode rexNode
@@ -360,7 +359,9 @@ public TimeColIdentifer()
     @Override
     public <R> R accept(SqlVisitor<R> visitor)
     {
-
+      // We overridde the "accept()" method, because the __time column's presence is determined when Calcite is converting
+      // the identifiers to the fully qualified column names with prefixes. This is where the validation exception can
+      // trigger
       try {
         return super.accept(visitor);
       }

sql/src/main/java/org/apache/druid/sql/calcite/aggregation/builtin/EarliestLatestBySqlAggregator.java

@@ -100,7 +100,6 @@ public Aggregation toDruidAggregation(
     }
 
     final String fieldName = EarliestLatestAnySqlAggregator.getColumnName(
-        plannerContext,
         virtualColumnRegistry,
         args.get(0),
         rexNodes.get(0)
@@ -113,7 +112,6 @@ public Aggregation toDruidAggregation(
             aggregatorName,
             fieldName,
             EarliestLatestAnySqlAggregator.getColumnName(
-                plannerContext,
                 virtualColumnRegistry,
                 args.get(1),
                 rexNodes.get(1)
@@ -140,7 +138,6 @@ public Aggregation toDruidAggregation(
             aggregatorName,
             fieldName,
             EarliestLatestAnySqlAggregator.getColumnName(
-                plannerContext,
                 virtualColumnRegistry,
                 args.get(1),
                 rexNodes.get(1)

sql/src/test/java/org/apache/druid/sql/calcite/CalciteQueryTest.java

@@ -638,8 +638,6 @@ public void testGroupBySingleColumnDescendingNoTopN()
   @Test
   public void testEarliestAggregators()
   {
-    msqIncompatible();
-
     testQuery(
         "SELECT "
         + "EARLIEST(cnt), EARLIEST(m1), EARLIEST(dim1, 10), EARLIEST(dim1, CAST(10 AS INTEGER)), "
@@ -1200,8 +1198,6 @@ public void testStringLatestByGroupByWithAlwaysFalseCondition()
   @Test
   public void testPrimitiveEarliestInSubquery()
   {
-    msqIncompatible();
-
     testQuery(
         "SELECT SUM(val1), SUM(val2), SUM(val3) FROM (SELECT dim2, EARLIEST(m1) AS val1, EARLIEST(cnt) AS val2, EARLIEST(m2) AS val3 FROM foo GROUP BY dim2)",
         ImmutableList.of(
@@ -1408,7 +1404,6 @@ public void testPrimitiveAnyInSubquery()
   @Test
   public void testStringEarliestSingleStringDim()
   {
-    msqIncompatible();
     testQuery(
         "SELECT dim2, EARLIEST(dim1,10) AS val FROM foo GROUP BY dim2",
         ImmutableList.of(
@@ -1524,8 +1519,6 @@ public void testStringAnyInSubquery()
   @Test
   public void testEarliestAggregatorsNumericNulls()
   {
-    msqIncompatible();
-
     testQuery(
         "SELECT EARLIEST(l1), EARLIEST(d1), EARLIEST(f1) FROM druid.numfoo",
         ImmutableList.of(
@@ -1583,8 +1576,6 @@ public void testLatestAggregatorsNumericNull()
   @Test
   public void testFirstLatestAggregatorsSkipNulls()
   {
-    msqIncompatible();
-
     final DimFilter filter;
     if (useDefault) {
       filter = notNull("dim1");
@@ -1697,8 +1688,6 @@ public void testAnyAggregatorsSkipNullsWithFilter()
   @Test
   public void testOrderByEarliestFloat()
   {
-    msqIncompatible();
-
     List<Object[]> expected;
     if (NullHandling.replaceWithDefault()) {
       expected = ImmutableList.of(
@@ -1744,8 +1733,6 @@ public void testOrderByEarliestFloat()
   @Test
   public void testOrderByEarliestDouble()
   {
-    msqIncompatible();
-
     List<Object[]> expected;
     if (NullHandling.replaceWithDefault()) {
       expected = ImmutableList.of(
@@ -1791,8 +1778,6 @@ public void testOrderByEarliestDouble()
   @Test
   public void testOrderByEarliestLong()
   {
-    msqIncompatible();
-
     List<Object[]> expected;
     if (NullHandling.replaceWithDefault()) {
       expected = ImmutableList.of(
@@ -9660,7 +9645,9 @@ public void testTimeseriesEmptyResultsAggregatorDefaultValues()
   @Test
   public void testTimeseriesEmptyResultsAggregatorDefaultValuesNonVectorized()
   {
+    // Empty-dataset aggregation queries in MSQ return an empty row, rather than a single row as SQL requires.
     msqIncompatible();
+
     cannotVectorize();
     skipVectorize();
     // timeseries with all granularity have a single group, so should return default results for given aggregators
@@ -9976,7 +9963,6 @@ public void testGroupByAggregatorDefaultValues()
   @Test
   public void testGroupByAggregatorDefaultValuesNonVectorized()
   {
-    msqIncompatible();
     cannotVectorize();
     skipVectorize();
     testQuery(