Allow parameter for ES|QL SAMPLE

Author: jan-elastic

x-pack/plugin/esql/src/main/antlr/EsqlBaseParser.g4

@@ -313,5 +313,5 @@ completionCommand
     ;
 
 sampleCommand
-    : DEV_SAMPLE probability=decimalValue
+    : DEV_SAMPLE probability=constant
     ;

x-pack/plugin/esql/src/main/java/org/elasticsearch/xpack/esql/action/EsqlCapabilities.java

@@ -1168,7 +1168,7 @@ public enum Cap {
         LUCENE_QUERY_EVALUATOR_QUERY_REWRITE,
 
         /**
-         * Support parameters for LiMIT command.
+         * Support parameters for LIMIT command.
          */
         PARAMETER_FOR_LIMIT,
 
@@ -1195,7 +1195,12 @@ public enum Cap {
         /**
          * Support knn function
          */
-        KNN_FUNCTION(Build.current().isSnapshot());
+        KNN_FUNCTION(Build.current().isSnapshot()),
+
+        /**
+         * Support parameters for SAMPLE command.
+         */
+        PARAMETER_FOR_SAMPLE(Build.current().isSnapshot());
 
         private final boolean enabled;
 

x-pack/plugin/esql/src/main/java/org/elasticsearch/xpack/esql/parser/EsqlBaseParser.interp

@@ -88,6 +88,7 @@
 import java.util.function.Function;
 
 import static java.util.Collections.emptyList;
+import static org.elasticsearch.xpack.esql.common.Failure.fail;
 import static org.elasticsearch.xpack.esql.core.type.DataType.KEYWORD;
 import static org.elasticsearch.xpack.esql.core.util.StringUtils.WILDCARD;
 import static org.elasticsearch.xpack.esql.expression.NamedExpressions.mergeOutputExpressions;
@@ -783,7 +784,15 @@ public Literal inferenceId(EsqlBaseParser.IdentifierOrParameterContext ctx) {
     }
 
     public PlanFactory visitSampleCommand(EsqlBaseParser.SampleCommandContext ctx) {
-        var probability = visitDecimalValue(ctx.probability);
-        return plan -> new Sample(source(ctx), probability, plan);
+        Source source = source(ctx);
+        Object val = expression(ctx.probability).fold(FoldContext.small() /* TODO remove me */);
+        if (val instanceof Double probability && probability > 0.0 && probability < 1.0) {
+            return input -> new Sample(source, new Literal(source, probability, DataType.DOUBLE), input);
+        } else {
+            throw new ParsingException(
+                source(ctx),
+                "invalid value for SAMPLE probability [" + val + "], expecting a number between 0 and 1, exclusive"
+            );
+        }
     }
 }

x-pack/plugin/esql/src/main/java/org/elasticsearch/xpack/esql/parser/EsqlBaseParser.java

@@ -162,6 +162,25 @@ public void testInvalidLimit() {
         assertEquals("1:13: Invalid value for LIMIT [-1], expecting a non negative integer", error("row a = 1 | limit -1"));
     }
 
+    public void testInvalidSample() {
+        assertEquals(
+            "1:13: invalid value for SAMPLE probability [foo], expecting a number between 0 and 1, exclusive",
+            error("row a = 1 | sample \"foo\"")
+        );
+        assertEquals(
+            "1:13: invalid value for SAMPLE probability [-1.0], expecting a number between 0 and 1, exclusive",
+            error("row a = 1 | sample -1.0")
+        );
+        assertEquals(
+            "1:13: invalid value for SAMPLE probability [0], expecting a number between 0 and 1, exclusive",
+            error("row a = 1 | sample 0")
+        );
+        assertEquals(
+            "1:13: invalid value for SAMPLE probability [1], expecting a number between 0 and 1, exclusive",
+            error("row a = 1 | sample 1")
+        );
+    }
+
     private String error(String query) {
         ParsingException e = expectThrows(ParsingException.class, () -> defaultAnalyzer.analyze(parser.createStatement(query)));
         String message = e.getMessage();

x-pack/plugin/esql/src/main/java/org/elasticsearch/xpack/esql/parser/LogicalPlanBuilder.java

@@ -104,6 +104,7 @@
 import static org.hamcrest.Matchers.hasSize;
 import static org.hamcrest.Matchers.instanceOf;
 import static org.hamcrest.Matchers.is;
+import static org.hamcrest.Matchers.startsWith;
 
 //@TestLogging(value = "org.elasticsearch.xpack.esql:TRACE", reason = "debug")
 public class StatementParserTests extends AbstractStatementParserTests {
@@ -3485,8 +3486,15 @@ public void testSample() {
         assumeTrue("SAMPLE requires corresponding capability", EsqlCapabilities.Cap.SAMPLE_V3.isEnabled());
         expectError("FROM test | SAMPLE .1 2", "line 1:23: extraneous input '2' expecting <EOF>");
         expectError("FROM test | SAMPLE .1 \"2\"", "line 1:23: extraneous input '\"2\"' expecting <EOF>");
-        expectError("FROM test | SAMPLE 1", "line 1:20: mismatched input '1' expecting {DECIMAL_LITERAL, '+', '-'}");
-        expectError("FROM test | SAMPLE", "line 1:19: mismatched input '<EOF>' expecting {DECIMAL_LITERAL, '+', '-'}");
+        expectError(
+            "FROM test | SAMPLE 1",
+            "line 1:13: invalid value for SAMPLE probability [1], expecting a number between 0 and 1, exclusive"
+        );
+        expectThrows(
+            ParsingException.class,
+            startsWith("line 1:19: mismatched input '<EOF>' expecting {"),
+            () -> statement("FROM test | SAMPLE")
+        );
     }
 
     static Alias alias(String name, Expression value) {

x-pack/plugin/esql/src/test/java/org/elasticsearch/xpack/esql/analysis/ParsingTests.java

@@ -455,14 +455,14 @@ FROM EVAL SORT LIMIT with documents_found:
         - method: POST
           path: /_query
           parameters: [ ]
-          capabilities: [ parameter_for_limit ]
+          capabilities: [ parameter_for_limit, parameter_for_sample ]
       reason: "named or positional parameters"
 
   - do:
       esql.query:
         body:
-          query: 'from test | eval x = ?, y = ?, z = ?, t = ?, u = ?, v = ? | keep x, y, z, t, u, v | limit ?'
-          params: [{"n1" : "1"}, {"n2" : 2.0}, {"n3" : null}, {"n4" : true}, {"n5" : 123}, {"n6": 1674835275193}, {"l": 3}]
+          query: 'from test | sample ? | eval x = ?, y = ?, z = ?, t = ?, u = ?, v = ? | keep x, y, z, t, u, v | limit ?'
+          params: [{"s": 0.99999999999}, {"n1" : "1"}, {"n2" : 2.0}, {"n3" : null}, {"n4" : true}, {"n5" : 123}, {"n6": 1674835275193}, {"l": 3}]
 
   - length: {columns: 6}
   - match: {columns.0.name: "x"}
@@ -519,14 +519,14 @@ FROM EVAL SORT LIMIT with documents_found:
         - method: POST
           path: /_query
           parameters: [ ]
-          capabilities: [ parameter_for_limit ]
+          capabilities: [ parameter_for_limit, parameter_for_sample ]
       reason: "named or positional parameters for field names"
 
   - do:
       esql.query:
         body:
-          query: 'from test | stats x = count(?f1), y = sum(?f2) by ?f3 | sort ?f3 | keep ?f3, x, y  | limit ?l'
-          params: [{"f1" : {"identifier" : "time"}}, {"f2" : { "identifier" : "count" }}, {"f3" : { "identifier" : "color"}}, {"l": 3}]
+          query: 'from test | sample ?s | stats x = count(?f1), y = sum(?f2) by ?f3 | sort ?f3 | keep ?f3, x, y  | limit ?l'
+          params: [{"s": 0.99999999999}, {"f1" : {"identifier" : "time"}}, {"f2" : { "identifier" : "count" }}, {"f3" : { "identifier" : "color"}}, {"l": 3}]
 
   - length: {columns: 3}
   - match: {columns.0.name: "color"}