marco-sulla / ewa Goto Github PK

Only if aaberg/sql2o#330 will be accepted.

in config.ini

case_sensitive = 1|0

if case sensitive, no lower() to attribute names

To Services, add an optional Connection parameter to the enrich() method.

Adds also the possibility to pass null as Connection parameter to Repository method. If null, method without Connection parameter is called.

Methods in repo with a parameter Connection should be allowed to get a null. If null, the corresponding method without the Connection parameter should be invoked.

Repository code generated with :

• Window 10
• Python 3.6.8
• database Microsoft sql Server
• java 11.0.3

Steps to reproduce:

1. The repository is generated as expected
2. An insert operation fails due cast exception

You can find the generated code here invitoRegistrazione.zip

The repository tries to return the inserted primary key (Long) but a cast to BigDecimal throws an exception executing res_true = ((BigDecimal) res).longValue();

Code under investigation :

        final Object res = Sql2oUtility.getInsertedId("INVITO_REGISTRAZIONE", "IUDINVITOREGISTRAZIONE", con, key, Long.class);

        Object res_true;
        final Class<?> klass = Long.class;
        
        if (res != null && (klass == Long.class || klass == BigDecimal.class)) {
            res_true = ((BigDecimal) res).longValue();
        }
        else {
            res_true = res;
        }

### Thanks from Niuma Epp Team

Enjoy!

Even if the code is auto-generated, is cumbersome that the same code is copy-pasted very much time.
I think it's better to implement a AbstractRepository and an Repository interface, that shares the common code to all repositories created with EWA. This abstract and interfacecan be useful also to people that do NOT want to use EWA. The same think should be done also for services.

For this reason, the findByNameOfUniqueId() must be renamed simply findById(). This also improves abstraction and encapsulation. Interface for models and aggregators should also be implemented.

PS: are specific interfaces for repo and services really needed? Evaluate to eliminate them.

in config.ini:

auto_derive_column_names = 1|0

If 1, in repository add
query.setAutoDeriveColumnNames(true)
to all methods

in model, use pyhumps

See also #8

The install instructions are really difficult. I think that:

1. A setup.py should be created
2. A wheel should be created
3. Release the last version with the wheel
4. publish it to Pypi

This way, installing EWA will be simply pip install ewa.

PS: obligatory requirement of mssql and cx_Oraclein code must be removed. The code that requires these libraries must be loaded only if the .ini file requires that database. There's no need to impose the user to install all the supported drivers.

PPS: instructions to how to install the absurdly compicated cx_Oracle driver should remain.

https://docs.oracle.com/javase/8/docs/api/java/sql/Connection.html#prepareStatement-java.lang.String-java.lang.String:A-

public Query createQuery(String queryText, String ... columnNames) {
    try {
        if (jdbcConnection.isClosed()) {
            createConnection();
        }
    } catch(SQLException e) {
        throw new Sql2oException("Error creating connection", e);
    }

    return new Query(this, queryText, columnNames);

}