A raw information is saved inside a database (multiple tables). it must be by hand checked and remedied. The checked data ought to be saved in database together with the raw data too. For the reason that situation, could it be wise decision to produce two separate databases (eg. raw_data and checked_data) ? Or there must be just one database? Thanks
In case your raw data and checked data will be very inormous than just use two different database
With normalization and taking advantage of methods you are able to keeping it in a single database.
There's no suggested method here except your personal preferences. You are able to keep cleaned data with raw data in same database however in different tables and might be prefix something similar to raw_ towards the raw data tables.
Otherwise you might have a seperate database for every kind of data. The advantages could be seperation while the drawback could be more expensive Join etc if have to be done between both of these.
In most cases it's a lot simpler to operate inside a single instance than across multiple instances. Distributed transactions perform reduced. They might require more typing (always needing to give a database link). This isn't only a matter of convenience but additionally of integrity. You might want to make sure that confirmed record is in both the RAW data set or even the Cleaned data set although not both. Checking this kind of factor is much more workable in one database.
The way you organize things in one database is dependent to some degree in your selected DBMS flavour, and what it really supports. You'll have a single schema (user account) and employ a naming convention for example prefix, for instance RAW_TABLE_1 and CLEAN_TABLE_1. Or you might want to use different schemas, which will help you to support the same table title, for instance RAW_USER.TABLE_1 and CLEAN_USER.TABLE_1. Both approaches have advantages. It is good to possess a constant indication of whether we're dealing with raw or clean data. However we might have tools or programs which we wish to use that expect the standard table names. Synonyms might help in connection with this.