ES|QL: No plain strings in Literal

x-pack/plugin/esql-core/src/main/java/org/elasticsearch/xpack/esql/core/expression/Literal.java

@@ -11,18 +11,24 @@
 import org.elasticsearch.common.io.stream.NamedWriteableRegistry;
 import org.elasticsearch.common.io.stream.StreamInput;
 import org.elasticsearch.common.io.stream.StreamOutput;
+import org.elasticsearch.common.lucene.BytesRefs;
 import org.elasticsearch.xpack.esql.core.QlIllegalArgumentException;
 import org.elasticsearch.xpack.esql.core.tree.NodeInfo;
 import org.elasticsearch.xpack.esql.core.tree.Source;
 import org.elasticsearch.xpack.esql.core.type.DataType;
 import org.elasticsearch.xpack.esql.core.util.PlanStreamInput;
+import org.elasticsearch.xpack.versionfield.Version;
 
 import java.io.IOException;
+import java.util.Collection;
 import java.util.List;
 import java.util.Objects;
 
 import static org.elasticsearch.xpack.esql.core.type.DataType.CARTESIAN_POINT;
 import static org.elasticsearch.xpack.esql.core.type.DataType.GEO_POINT;
+import static org.elasticsearch.xpack.esql.core.type.DataType.KEYWORD;
+import static org.elasticsearch.xpack.esql.core.type.DataType.TEXT;
+import static org.elasticsearch.xpack.esql.core.type.DataType.VERSION;
 import static org.elasticsearch.xpack.esql.core.util.SpatialCoordinateTypes.CARTESIAN;
 import static org.elasticsearch.xpack.esql.core.util.SpatialCoordinateTypes.GEO;
 
@@ -45,10 +51,25 @@ public class Literal extends LeafExpression {
 
     public Literal(Source source, Object value, DataType dataType) {
         super(source);
+        assert noPlainStrings(value, dataType);
         this.dataType = dataType;
         this.value = value;
     }
 
+    private boolean noPlainStrings(Object value, DataType dataType) {
+        if (dataType == KEYWORD || dataType == TEXT || dataType == VERSION) {
+            if (value == null) {
+                return true;
+            }
+            return switch (value) {
+                case String s -> false;
+                case Collection<?> c -> c.stream().allMatch(x -> noPlainStrings(x, dataType));
+                default -> true;
+            };
+        }
+        return true;
+    }
+
     private static Literal readFrom(StreamInput in) throws IOException {
         Source source = Source.readFrom((StreamInput & PlanStreamInput) in);
         Object value = in.readGenericValue();
@@ -122,7 +143,21 @@ public boolean equals(Object obj) {
 
     @Override
     public String toString() {
-        String str = String.valueOf(value);
+        String str;
+        if (dataType == KEYWORD || dataType == TEXT) {
+            str = BytesRefs.toString(value);
+        } else if (dataType == VERSION && value instanceof BytesRef br) {
+            str = new Version(br).toString();
+            // TODO review how we manage IPs: https://github.com/elastic/elasticsearch/issues/129605
+            // } else if (dataType == IP && value instanceof BytesRef ip) {
+            // str = DocValueFormat.IP.format(ip);
+        } else {
+            str = String.valueOf(value);
+        }
+
+        if (str == null) {
+            str = "null";
+        }
         if (str.length() > 500) {
             return str.substring(0, 500) + "...";
         }
@@ -154,6 +189,10 @@ public static Literal of(Expression source, Object value) {
         return new Literal(source.source(), value, source.dataType());
     }
 
+    public static Literal keyword(Source source, String literal) {
+        return new Literal(source, BytesRefs.toBytesRef(literal), KEYWORD);
+    }
+
     /**
      * Not all literal values are currently supported in StreamInput/StreamOutput as generic values.
      * This mapper allows for addition of new and interesting values without (yet) adding to StreamInput/Output.

x-pack/plugin/esql-core/src/main/java/org/elasticsearch/xpack/esql/core/expression/MapExpression.java

@@ -10,6 +10,7 @@
 import org.elasticsearch.common.io.stream.NamedWriteableRegistry;
 import org.elasticsearch.common.io.stream.StreamInput;
 import org.elasticsearch.common.io.stream.StreamOutput;
+import org.elasticsearch.common.lucene.BytesRefs;
 import org.elasticsearch.xpack.esql.core.tree.NodeInfo;
 import org.elasticsearch.xpack.esql.core.tree.Source;
 import org.elasticsearch.xpack.esql.core.type.DataType;
@@ -39,7 +40,7 @@ public class MapExpression extends Expression {
 
     private final Map<Expression, Expression> map;
 
-    private final Map<Object, Expression> keyFoldedMap;
+    private final Map<String, Expression> keyFoldedMap;
 
     public MapExpression(Source source, List<Expression> entries) {
         super(source, entries);
@@ -54,7 +55,7 @@ public MapExpression(Source source, List<Expression> entries) {
             entryExpressions.add(new EntryExpression(key.source(), key, value));
             map.put(key, value);
             if (key instanceof Literal l) {
-                this.keyFoldedMap.put(l.value(), value);
+                this.keyFoldedMap.put(BytesRefs.toString(l.value()), value);
             }
         }
     }
@@ -95,7 +96,7 @@ public Map<Expression, Expression> map() {
         return map;
     }
 
-    public Map<Object, Expression> keyFoldedMap() {
+    public Map<String, Expression> keyFoldedMap() {
         return keyFoldedMap;
     }
 

x-pack/plugin/esql-core/src/test/java/org/elasticsearch/xpack/esql/core/expression/LiteralTests.java

@@ -8,6 +8,7 @@
 
 import joptsimple.internal.Strings;
 
+import org.elasticsearch.common.lucene.BytesRefs;
 import org.elasticsearch.test.ESTestCase;
 import org.elasticsearch.xpack.esql.core.InvalidArgumentException;
 import org.elasticsearch.xpack.esql.core.tree.AbstractNodeTestCase;
@@ -20,7 +21,6 @@
 import java.util.ArrayList;
 import java.util.Arrays;
 import java.util.List;
-import java.util.Objects;
 import java.util.function.Function;
 import java.util.function.Supplier;
 
@@ -62,7 +62,7 @@ static class ValueAndCompatibleTypes {
         new ValueAndCompatibleTypes(ESTestCase::randomLong, LONG, FLOAT, DOUBLE, BOOLEAN),
         new ValueAndCompatibleTypes(ESTestCase::randomFloat, FLOAT, LONG, DOUBLE, BOOLEAN),
         new ValueAndCompatibleTypes(ESTestCase::randomDouble, DOUBLE, LONG, FLOAT, BOOLEAN),
-        new ValueAndCompatibleTypes(() -> randomAlphaOfLength(5), KEYWORD)
+        new ValueAndCompatibleTypes(() -> BytesRefs.toBytesRef(randomAlphaOfLength(5)), KEYWORD)
     );
 
     public static Literal randomLiteral() {
@@ -129,14 +129,14 @@ public void testReplaceChildren() {
 
     public void testToString() {
         assertThat(new Literal(Source.EMPTY, 1, LONG).toString(), equalTo("1"));
-        assertThat(new Literal(Source.EMPTY, "short", KEYWORD).toString(), equalTo("short"));
+        assertThat(new Literal(Source.EMPTY, BytesRefs.toBytesRef("short"), KEYWORD).toString(), equalTo("short"));
         // toString should limit it's length
         String tooLong = Strings.repeat('a', 510);
-        assertThat(new Literal(Source.EMPTY, tooLong, KEYWORD).toString(), equalTo(Strings.repeat('a', 500) + "..."));
+        assertThat(new Literal(Source.EMPTY, BytesRefs.toBytesRef(tooLong), KEYWORD).toString(), equalTo(Strings.repeat('a', 500) + "..."));
 
         for (ValueAndCompatibleTypes g : GENERATORS) {
             Literal lit = new Literal(Source.EMPTY, g.valueSupplier.get(), randomFrom(g.validDataTypes));
-            assertThat(lit.toString(), equalTo(Objects.toString(lit.value())));
+            assertThat(lit.toString(), equalTo(BytesRefs.toString(lit.value())));
         }
     }
 

x-pack/plugin/esql-core/src/test/java/org/elasticsearch/xpack/esql/core/expression/function/scalar/FunctionTestUtils.java

@@ -7,6 +7,7 @@
 
 package org.elasticsearch.xpack.esql.core.expression.function.scalar;
 
+import org.elasticsearch.common.lucene.BytesRefs;
 import org.elasticsearch.xpack.esql.core.expression.Literal;
 import org.elasticsearch.xpack.esql.core.type.DataType;
 
@@ -15,10 +16,13 @@
 public final class FunctionTestUtils {
 
     public static Literal l(Object value) {
-        return new Literal(EMPTY, value, DataType.fromJava(value));
+        return l(value, DataType.fromJava(value));
     }
 
     public static Literal l(Object value, DataType type) {
+        if ((type == DataType.TEXT || type == DataType.KEYWORD) && value instanceof String) {
+            value = BytesRefs.toBytesRef(value);
+        }
         return new Literal(EMPTY, value, type);
     }
 }

x-pack/plugin/esql-core/src/test/java/org/elasticsearch/xpack/esql/core/util/TestUtils.java

@@ -7,6 +7,7 @@
 
 package org.elasticsearch.xpack.esql.core.util;
 
+import org.elasticsearch.common.lucene.BytesRefs;
 import org.elasticsearch.xpack.esql.core.expression.FieldAttribute;
 import org.elasticsearch.xpack.esql.core.expression.Literal;
 import org.elasticsearch.xpack.esql.core.tree.Source;
@@ -22,6 +23,8 @@
 import static org.elasticsearch.test.ESTestCase.randomFrom;
 import static org.elasticsearch.xpack.esql.core.tree.Source.EMPTY;
 import static org.elasticsearch.xpack.esql.core.type.DataType.INTEGER;
+import static org.elasticsearch.xpack.esql.core.type.DataType.KEYWORD;
+import static org.elasticsearch.xpack.esql.core.type.DataType.TEXT;
 
 public final class TestUtils {
     private TestUtils() {}
@@ -39,6 +42,10 @@ public static Literal of(Source source, Object value) {
         if (value instanceof Literal) {
             return (Literal) value;
         }
+        DataType type = DataType.fromJava(value);
+        if ((type == TEXT || type == KEYWORD) && value instanceof String) {
+            value = BytesRefs.toBytesRef(value);
+        }
         return new Literal(source, value, DataType.fromJava(value));
     }
 
@@ -58,12 +65,16 @@ public static FieldAttribute getFieldAttribute(String name, DataType dataType) {
         return new FieldAttribute(EMPTY, name, new EsField(name + "f", dataType, emptyMap(), true));
     }
 
-    /** Similar to {@link String#strip()}, but removes the WS throughout the entire string. */
+    /**
+     * Similar to {@link String#strip()}, but removes the WS throughout the entire string.
+     */
     public static String stripThrough(String input) {
         return WS_PATTERN.matcher(input).replaceAll(StringUtils.EMPTY);
     }
 
-    /** Returns the input string, but with parts of it having the letter casing changed. */
+    /**
+     * Returns the input string, but with parts of it having the letter casing changed.
+     */
     public static String randomCasing(String input) {
         StringBuilder sb = new StringBuilder(input.length());
         for (int i = 0, inputLen = input.length(), step = (int) Math.sqrt(inputLen); i < inputLen; i += step) {

x-pack/plugin/esql/qa/testFixtures/src/main/java/org/elasticsearch/xpack/esql/EsqlTestUtils.java

@@ -21,6 +21,7 @@
 import org.elasticsearch.common.breaker.CircuitBreaker;
 import org.elasticsearch.common.breaker.NoopCircuitBreaker;
 import org.elasticsearch.common.bytes.BytesReference;
+import org.elasticsearch.common.lucene.BytesRefs;
 import org.elasticsearch.common.regex.Regex;
 import org.elasticsearch.common.settings.Settings;
 import org.elasticsearch.common.util.BigArrays;
@@ -227,7 +228,11 @@ public static Literal of(Source source, Object value) {
         if (value instanceof Literal) {
             return (Literal) value;
         }
-        return new Literal(source, value, DataType.fromJava(value));
+        var dataType = DataType.fromJava(value);
+        if (value instanceof String) {
+            value = BytesRefs.toBytesRef(value);
+        }
+        return new Literal(source, value, dataType);
     }
 
     public static ReferenceAttribute referenceAttribute(String name, DataType type) {

x-pack/plugin/esql/qa/testFixtures/src/main/resources/ip.csv-spec

@@ -398,7 +398,7 @@ eth0           |gamma          |fe80::cae2:65ff:fece:feb9
 lo0            |gamma          |fe80::cae2:65ff:fece:feb9                                                         |fe81::cae2:65ff:fece:feb9
 ;
 
-cdirMatchEqualsInsOrsIPs
+cidrMatchEqualsInsOrsIPs
 required_capability: combine_disjunctive_cidrmatches
 
 FROM hosts

x-pack/plugin/esql/qa/testFixtures/src/main/resources/match-function.csv-spec

@@ -860,3 +860,21 @@ c: long     |   scalerank: long
 10             | 3              
 15             | 2              
 ;
+
+
+testMatchWithReplace
+required_capability: match_function
+required_capability: no_plain_strings_in_literals
+from books 
+| keep book_no, author 
+| where match(author, REPLACE("FaulkneX", "X", "r"))
+| sort book_no 
+| limit 5;
+
+book_no:keyword | author:text
+2378            | [Carol Faulkner, Holly Byers Ochoa, Lucretia Mott]
+2713            | William Faulkner
+2847            | Colleen Faulkner
+2883            | William Faulkner
+3293            | Danny Faulkner
+;

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

@@ -1207,7 +1207,15 @@ public enum Cap {
         /**
          * Support parameters for SAMPLE command.
          */
-        PARAMETER_FOR_SAMPLE(Build.current().isSnapshot());
+        PARAMETER_FOR_SAMPLE(Build.current().isSnapshot()),
+
+        /**
+         * From now, Literal only accepts strings as BytesRefs.
+         * No java.lang.String anymore.
+         *
+         * https://github.com/elastic/elasticsearch/issues/129322
+         */
+        NO_PLAIN_STRINGS_IN_LITERALS;
 
         private final boolean enabled;
 

x-pack/plugin/esql/src/main/java/org/elasticsearch/xpack/esql/analysis/Analyzer.java

@@ -9,6 +9,7 @@
 
 import org.elasticsearch.common.logging.HeaderWarning;
 import org.elasticsearch.common.logging.LoggerMessageFormat;
+import org.elasticsearch.common.lucene.BytesRefs;
 import org.elasticsearch.compute.data.Block;
 import org.elasticsearch.core.Strings;
 import org.elasticsearch.index.IndexMode;
@@ -318,7 +319,7 @@ protected LogicalPlan rule(Enrich plan, AnalyzerContext context) {
                 // the policy does not exist
                 return plan;
             }
-            final String policyName = (String) plan.policyName().fold(FoldContext.small() /* TODO remove me */);
+            final String policyName = BytesRefs.toString(plan.policyName().fold(FoldContext.small() /* TODO remove me */));
             final var resolved = context.enrichResolution().getResolvedPolicy(policyName, plan.mode());
             if (resolved != null) {
                 var policy = new EnrichPolicy(resolved.matchType(), null, List.of(), resolved.matchField(), resolved.enrichFields());
@@ -401,7 +402,7 @@ private static class ResolveInference extends ParameterizedAnalyzerRule<Inferenc
         protected LogicalPlan rule(InferencePlan<?> plan, AnalyzerContext context) {
             assert plan.inferenceId().resolved() && plan.inferenceId().foldable();
 
-            String inferenceId = plan.inferenceId().fold(FoldContext.small()).toString();
+            String inferenceId = BytesRefs.toString(plan.inferenceId().fold(FoldContext.small()));
             ResolvedInference resolvedInference = context.inferenceResolution().getResolvedInference(inferenceId);
 
             if (resolvedInference != null && resolvedInference.taskType() == plan.taskType()) {
@@ -431,7 +432,7 @@ protected LogicalPlan rule(Lookup lookup, AnalyzerContext context) {
             // the parser passes the string wrapped in a literal
             Source source = lookup.source();
             Expression tableNameExpression = lookup.tableName();
-            String tableName = lookup.tableName().toString();
+            String tableName = BytesRefs.toString(tableNameExpression.fold(FoldContext.small() /* TODO remove me */));
             Map<String, Map<String, Column>> tables = context.configuration().tables();
             LocalRelation localRelation = null;
 
@@ -1563,18 +1564,22 @@ private static boolean supportsStringImplicitCasting(DataType type) {
         }
 
         private static UnresolvedAttribute unresolvedAttribute(Expression value, String type, Exception e) {
+            String name = BytesRefs.toString(value.fold(FoldContext.small()) /* TODO remove me */);
             String message = LoggerMessageFormat.format(
+                null,
                 "Cannot convert string [{}] to [{}], error [{}]",
-                value.fold(FoldContext.small() /* TODO remove me */),
+                name,
                 type,
                 (e instanceof ParsingException pe) ? pe.getErrorMessage() : e.getMessage()
             );
-            return new UnresolvedAttribute(value.source(), String.valueOf(value.fold(FoldContext.small() /* TODO remove me */)), message);
+            return new UnresolvedAttribute(value.source(), name, message);
         }
 
         private static Expression castStringLiteralToTemporalAmount(Expression from) {
             try {
-                TemporalAmount result = maybeParseTemporalAmount(from.fold(FoldContext.small() /* TODO remove me */).toString().strip());
+                TemporalAmount result = maybeParseTemporalAmount(
+                    BytesRefs.toString(from.fold(FoldContext.small() /* TODO remove me */)).strip()
+                );
                 if (result == null) {
                     return from;
                 }

x-pack/plugin/esql/src/main/java/org/elasticsearch/xpack/esql/expression/function/aggregate/Count.java

@@ -163,7 +163,7 @@ public Expression surrogate() {
             return new Mul(
                 s,
                 new Coalesce(s, new MvCount(s, field), List.of(new Literal(s, 0, DataType.INTEGER))),
-                new Count(s, new Literal(s, StringUtils.WILDCARD, DataType.KEYWORD))
+                new Count(s, Literal.keyword(s, StringUtils.WILDCARD))
             );
         }
 

x-pack/plugin/esql/src/main/java/org/elasticsearch/xpack/esql/expression/function/aggregate/Sum.java

@@ -143,8 +143,6 @@ public Expression surrogate() {
         }
 
         // SUM(const) is equivalent to MV_SUM(const)*COUNT(*).
-        return field.foldable()
-            ? new Mul(s, new MvSum(s, field), new Count(s, new Literal(s, StringUtils.WILDCARD, DataType.KEYWORD)))
-            : null;
+        return field.foldable() ? new Mul(s, new MvSum(s, field), new Count(s, Literal.keyword(s, StringUtils.WILDCARD))) : null;
     }
 }