I am using Spring Data support for Elasticsearch. Here is the timestamp field mapping:
@Field(type = FieldType.Date, index = FieldIndex.not_analyzed, store =
Check https://github.com/spring-projects/spring-data-elasticsearch/wiki/Custom-ObjectMapper to add JavaTimeModule
to your ObjectMapper
.
@Configuration
public class ElasticSearchConfiguration {
@Bean
public ElasticsearchTemplate elasticsearchTemplate(Client client) {
return new ElasticsearchTemplate(client, new CustomEntityMapper());
}
public static class CustomEntityMapper implements EntityMapper {
private final ObjectMapper objectMapper;
public CustomEntityMapper() {
objectMapper = new ObjectMapper();
objectMapper.configure(DeserializationFeature.FAIL_ON_UNKNOWN_PROPERTIES, false);
objectMapper.configure(DeserializationFeature.ACCEPT_SINGLE_VALUE_AS_ARRAY, true);
objectMapper.registerModule(new CustomGeoModule());
objectMapper.registerModule(new JavaTimeModule());
}
@Override
public String mapToString(Object object) throws IOException {
return objectMapper.writeValueAsString(object);
}
@Override
public <T> T mapToObject(String source, Class<T> clazz) throws IOException {
return objectMapper.readValue(source, clazz);
}
}
}
I was experiencing the similar issue: 'Z' in the date value is treated as a character, so the date parse failed. My solution is using custom pattern to make sure the conversion correct:
@Field(type = FieldType.Date, format = DateFormat.custom, pattern = "yyyy-MM-dd'T'HH:mm:ss'Z'")
private LocalDateTime dateField;
And if we don't want to repeat the pattern on different fields again and again, we can try to centralize to the conversion logic. Spring Data Elastic Search provide the custom conversion feature, check here for the example.
Basically we can write a converter from String to LocalDateTime, and put the date pattern over there.