SuperCSV Dozer deep-mapping map with enum and subclasses?

Question

I'm trying to see if this is even possible with SuperCSV and Dozer, or if I should just revert to Map parsing. I have a POJO that has a member field of Map. Thankfully, during CSV parsing, I know what specific subclass of MyInterface should be built, and also the value of MyEnum will be static. But how would I set all this up in the column mappings? Thanks!

Currently, my cell processors have this structure, and I am using a CsvMapReader.

private static final CellProcessor[] CELL_PROCESSORS = new CellProcessor[] {
        new NotNull(new Trim(new StrRegEx("^\\d{10,}$"))),  // phone1
        new Optional(new Trim(new StrRegEx("^\\d{10,}$"))), // phone2
        new Optional(new Trim(new StrRegEx("^\\d{10,}$"))), // phone3
        new Optional(new Trim()),                           // callVar1
        new Optional(new Trim()),                           // callVar2
        new Optional(new Trim()),                           // callVar3
        new Optional(new Trim()),                           // callVar4
        new Optional(new Trim()),                           // callVar5
        new Optional(new Trim()),                           // callVar6
        new Optional(new Trim()),                           // callVar7
        new Optional(new Trim()),                           // callVar8
        new Optional(new Trim()),                           // callVar9
        new Optional(new Trim()),                           // callVar10
};

private Contact mapRowToContact(Map<String, Object> row) {
    Contact contact = new Contact();

    MyPhoneContactMethodData methodData = new MyPhoneContactMethodData();

    List<Phone> phones = new ArrayList<>();
    Phone phone = new Phone();
    phone.setPhoneNumber((String)row.get("phone1"));
    phones.add(phone);
    phone = new Phone();
    phone.setPhoneNumber((String)row.get("phone2"));
    if (phone.getPhoneNumber() != null) {
        phones.add(phone);
    }
    phone = new Phone();
    phone.setPhoneNumber((String)row.get("phone3"));
    if (phone.getPhoneNumber() != null) {
        phones.add(phone);
    }
    methodData.setPhones(phones);

    List<String> callVars = new ArrayList<>();
    callVars.add((String)row.get("callVar1"));
    callVars.add((String)row.get("callVar2"));
    callVars.add((String)row.get("callVar3"));
    callVars.add((String)row.get("callVar4"));
    callVars.add((String)row.get("callVar5"));
    callVars.add((String)row.get("callVar6"));
    callVars.add((String)row.get("callVar7"));
    callVars.add((String)row.get("callVar8"));
    callVars.add((String)row.get("callVar9"));
    callVars.add((String)row.get("callVar10"));
    methodData.setEnterpriseCallVarData(callVars);

    Map<ContactMethod, ContactMethodData> methodDataMap = new HashMap<>();
    methodDataMap.put(ContactMethod.PHONE, methodData);
    contact.setContactMethodData(methodDataMap);

    return contact;
}

A Contact has this structure, with many other unrelated fields:

public class Contact {
    private Integer id;
    private Map<ContactMethod, ContactMethodData> contactMethodData;
}

ContactMethod is an enum, with values PHONE and EMAIL. ContactMethodData is an interface, of which the superclass of MyPhoneContactMethodData implements.

Can you post some example code? I'm a Super CSV dev - I can help, but only if I understand what the problem is! :) — James Bassett, Feb 14 '13 at 22:24

James Bassett · Accepted Answer · 2013-02-20T23:14:21.397

Thanks for the code - much easier to understand now :)

You should be able to read each row of CSV as a MyPhoneContactMethodData instance by using the following bean mapping. Just make sure you call configureBeanMapping() with this before reading (as shown on the Super CSV website).

You will then have to manually create the Contact and add the MyPhoneContactMethodData to the contactMethodData map, using with ContactMethod.PHONE as the key (as done in the last 3 lines of your code).

final String[] beanMapping = new String[]{
    "phones[0].phoneNumber",
    "phones[1].phoneNumber",
    "phones[2].phoneNumber",
    "enterpriseCallVarData[0]",
    "enterpriseCallVarData[1]",
    "enterpriseCallVarData[2]",
    "enterpriseCallVarData[3]",
    "enterpriseCallVarData[4]",
    "enterpriseCallVarData[5]",
    "enterpriseCallVarData[6]",
    "enterpriseCallVarData[7]",
    "enterpriseCallVarData[8]",
    "enterpriseCallVarData[9]"
};

beanReader.configureBeanMapping(MyPhoneContactMethodData.class, beanMapping);

MyPhoneContactMethodData methodData;
while( (methodData = 
    beanReader.read(MyPhoneContactMethodData.class, CELL_PROCESSORS)) != null ) {
    // add to contact
}

Thanks. That makes sense. I may end up sticking with CsvMapReader in the end anyway, due to how complex some of the other fields in Contact are, which were not shown here in order to keep things isolated to the question at hand. — Scott Balmos, Feb 21 '13 at 14:49

SuperCSV Dozer deep-mapping map with enum and subclasses?

1 Answers1