[jira] [Commented] (CSV-305) Add data connector

["Gary D. Gregory (Jira)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/CSV-305?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17614457#comment-17614457 ] 

Gary D. Gregory commented on CSV-305:
-------------------------------------

Hi [~yotam2010] 

If you want an ORM API, you should look at JPA, as I do not think anything we provide here should be anything that would inevitably grow into an ORM framework, no matter how initially limited. You could also look at Apache Commons DBUtils. IMO Commons CSV should be only about reading and writing CSV files, not mapping to classes, there is no end to features for that, see also Commons BeanUtils.

> Add data connector
> ------------------
>
>                 Key: CSV-305
>                 URL: https://issues.apache.org/jira/browse/CSV-305
>             Project: Commons CSV
>          Issue Type: Improvement
>            Reporter: Yotam Hadas
>            Priority: Minor
>
> After using the csv library I felt like I had 1 main thing missing - something to connect between the "infrastructure" (CSVFormat / CSVPrinter) and the actual data.
> Using "setHeader" / "printHeaders" and "printRecord" feels weird:
> 1. Order matters, when adding more data to the CSV you can easily write values in the wrong column.
> 2. It will make it easier to understand from the code what the CSV result should look like.
> Some *kotlin* code I wrote to achieve this partially in a project I work on:
> Helper classes:
> {code:java}
> class CSVData() {
>     private val columns: MutableList<CSVColumn> = mutableListOf()
>     fun getHeaders() = columns.map(CSVColumn::header).toTypedArray()
>     fun getRowValues(objectToConvert: Any) = columns.map(CSVColumn::columnValueProvider)
>         .map { columnValueProvider -> columnValueProvider.invoke(objectToConvert) }
>         .toTypedArray()
>     fun addColumn(column: CSVColumn) {
>         columns.add(column)
>     }
> }
> data class CSVColumn(val header: String, val columnValueProvider: (T: Any) -> Any?) {code}
> How it is used:
> {code:java}
> val csvData = CSVData() 
> csvData.addColumn(CSVColumn("First name"){(it as User).firstName}) 
> csvData.addColumn(CSVColumn("Last name"){(it as User).lastName})
> csvData.addColumn(CSVColumn("Payment information type") { (it as User).paymentInformation?.type })
> CSVPrinter(
>     FileWriter(tempFile),
>     CSVFormat.DEFAULT
>         .builder()
>         .setHeader(
>             *csvData.getHeaders()
>         ).build()
> ).use {
>     for (user in users) {
>         it.printRecord(
>             *csvData.getRowValues(user)
>         )
>     }
>     it.flush()
> }{code}
> What is missing - currently it works with only 1 record type.
> A better solution will enable to get the values from multiple objects like:
> {code:java}
> *csvData.getRowValues(user,userAdditionalDetails) {code}



--
This message was sent by Atlassian Jira
(v8.20.10#820010)