Type Support

The table recommends Java property types for common Elasticsearch fields. See Java/JDBC types.

Type Mappings

Elasticsearch field type	Java type	Description
keyword	String	Read text by field name or entity property mapping.
text	String	Read text by field name or entity property mapping.
integer	Integer	Signed 32-bit integer.
long	Long	Signed 64-bit integer.
double	Double	64-bit floating-point value.
boolean	Boolean	Use Boolean semantics, not a general integer property.
date	java.sql.Date (date-only use)	Does not preserve instant millisecond precision.
object	Map / List / Bean	See JSON Field Mapping.
keyword	Enum	Store enum names; declare the concrete enum class and keep field values aligned with constant names.

Text Length

text and keyword do not provide a VARCHAR(100)-style write-length constraint. dbVisitor does not reject values longer than 100 characters on their behalf; validate business length limits before writing.

The keyword setting ignore_above controls indexing, not write rejection. With ignore_above: 100, a 101-character value remains in _source, but that field is excluded from exact-match searches and aggregations. ignore_above reference

Example: Bind a Boolean and Read Its Field

jdbcTemplate.execute("PUT /type_example");
jdbcTemplate.executeUpdate("POST /type_example/_doc {\"id\": ?, \"enabled\": ?}",
        new Object[] { 1, true });
jdbcTemplate.execute("POST /type_example/_refresh");
Boolean enabled = jdbcTemplate.queryForObject(
        "POST /type_example/_search {\"query\": {\"term\": {\"id\": ?}}}",
        new Object[] { 1 }, (rs, rowNum) -> rs.getBoolean("enabled"));

Limits

• Date mapping does not preserve millisecond precision.
• BigInteger, exact decimals, and arbitrary binary content are not guaranteed to round-trip losslessly.
• Read business fields by name, RowMapper, or entity; the first two columns are _ID and _DOC.
• Applies to the ES6 and ES7 adapters.

Array Types

Elasticsearch does not declare a separate array type: a field mapped as integer, float, or keyword can hold multiple values. The adapter supports JDBC ARRAY binding and reads for these values. Use Integer[], Float[], or String[] properties; keep all values compatible with the field mapping.

Use an empty array_example index with id and int_array mapped as integer:

import java.sql.Types;
import net.hasor.dbvisitor.types.SqlArg;

Integer[] values = { 10, 20, 30 };
jdbc.executeUpdate("POST /array_example/_doc {\"id\": ?, \"int_array\": ?}",
        new Object[] { 1, SqlArg.valueOf(values, Types.ARRAY) });
jdbc.execute("POST /array_example/_refresh");
Integer[] loaded = jdbc.queryForObject(
        "POST /array_example/_search {\"_source\": [\"int_array\"], "
                + "\"query\": {\"term\": {\"id\": ?}}}",
        new Object[] { 1 }, Integer[].class);