Three questions on importing data

I am importing csv format data exported from Salesforce into Dynamics, the records have an 'ID' field which is the "ID" in Salesforce.

I accidentally mapped this field to the ID field(Primary Key) in Dynamics.

So I have 3 questions now:

1. Should I create a SFDC_ID field to store that ID field in source file and ignore the ID field in Dynamics while mapping(Let system generates IDs itself)

2. About duplicate detection:

I have started to delete those records with SFDC_ID as their primary ID. However, I imported 400000 Account records and they have been deleting every slow.

Can I start a new import to import the same records using  a new mapping file(ignore ID field)?

3. If the import wizard creates text field for me, what is the length of the text  field?

Appreciate your help.