fix(schema): Add duplicate field ID checking mechanism

[dttung2905]

As outlined in the issue, currently we do not perform field id duplication check in this IndexNameByID method. This leads to the issue of last one wins if multiple field ID are presented in the schema

iceberg-go/schema.go

Lines 788 to 800 in f886a24

	func IndexNameByID(schema *Schema) (map[int]string, error) {
	indexer := &indexByName{
	index: make(map[string]int),
	shortNameId: make(map[string]int),
	fieldNames: make([]string, 0),
	shortFieldNames: make([]string, 0),
	}
	if _, err := Visit(schema, indexer); err != nil {
	return nil, err
	}
	return indexer.ByID(), nil
	}

iceberg-go/schema.go

Lines 810 to 817 in f886a24

	func (i *indexByName) ByID() map[int]string {
	idToName := make(map[int]string)
	for k, v := range i.index {
	idToName[v] = k
	}
	return idToName
	}

I added a quick validation under ByID() and panic if duplicates are found and put this schema validation under init() as suggested in the issue

Fixes #593

[twuebi]

I think we should not have runtime panics that can be caused by user error, we should make the schema init fallible instead and return an err

[zeroshade]

why change the field name? Isn't this intended to test the case of multiple fields with the same name but different IDs?

[zeroshade]

I prefer that we forbid duplicate IDs in the first place so that we don't need to have this check at all

[dttung2905]

Let me find another way to go around this . Are you saying we test it in the addField() method? 🤔

[zeroshade]

that would be a good spot yea

[dttung2905]

This is ballooning into something more complicated than I thought initally. Let me find some time this weekend to look thoroughly into this