I have written a SpringBoot app that consumes a rest api and presents a rest api. My model pojo's have camelCase named properties. The json that the app consumes has under_score property names. The json that the app produces has under_score property names. I would like to use a PropertyNamingStrategy that will do the conversion automatically between Java and json names during marshalling/unmarshalling.
I have Java config that attempts to set a naming strategy that can handle this;
/**
* Configuration for Rest api.
* <p>
* Created by emurphy on 2/25/16.
*/
@Configuration
public class RestConfig
{
/**
* Bean to make jackson automatically convert from
* camelCase (java) to under_scores (json) in property names
*
* @return ObjectMapper that maps from Java camelCase to json under_score names
*/
@Bean
public ObjectMapper jacksonObjectMapper()
{
return new ObjectMapper().setPropertyNamingStrategy(new UpperCaseUnderscoreStrategy());
}
/**
* Property naming strategy that converts both ways between camelCase and under_score
* property names.
*/
public static class UpperCaseUnderscoreStrategy extends PropertyNamingStrategy.PropertyNamingStrategyBase
{
/**
* Converts camelCase to under_score and
* visa versa. The idea is that this
* name strategy can be used for both
* marshalling and unmarshaling.
*
* For example, "userName" would be converted to
* "user_name" and conversely "user_name" would
* be converted to "userName".
*
* @param input formatted as camelCase or under_score string
* @return input converted to opposite format
*/
@Override
public String translate(String input)
{
if (input == null || input.length() == 0)
{
return input; // garbage in, garbage out
}
//
// we always take the first character;
// this preserves initial underscore
//
StringBuilder sb = new StringBuilder();
final int length = input.length();
int i = 0;
//
// skip initial underscores
//
while ((i < length) && ('_' == input.charAt(i)))
{
sb.append(input.charAt(i));
i += 1;
}
while (i < length)
{
//
// find underscores, remove and capitalize next letter
//
while ((i < length) && ('_' != input.charAt(i)) && !Character.isUpperCase(input.charAt(i)))
{
sb.append(input.charAt(i));
i += 1;
}
if(i < length)
{
if('_' == input.charAt(i))
{
// underscore to uppercase
//
// skip underscores
//
while ((i < length) && ('_' == input.charAt(i)))
{
// skip underscores
i += 1;
}
//
// capitalize
//
if (i < length)
{
sb.append(Character.toUpperCase(input.charAt(i)));
i += 1;
}
}
else // uppercase to unscore + lowercase
{
sb.append('_');
sb.append(Character.toLowerCase(input.charAt(i)));
i += 1;
}
}
}
return sb.toString();
}
}
I can see the naming strategy's translate method getting called when my rest service converts Java pojos to json for the response. However, when I'm consuming a rest api, via RestTemplate, I don't see this get called and my resulting pojos are not correctly intialized from the incoming json; all properties whose name would need translation are null. This has forced me to use @JsonProperty on most properties. I have a lot of properties and a lot of pojos - this is a very inelegant, boilerplate kind of solution that SpringBoot is supposed to help with. Is there a way I can set a PropertyNamingStrategy that RestTemplate will use to convert the incoming json names from under_score to camelCase?
Thanks for your help.