It is a concept we implement in DWHsing where we use it to load data from differ

ent kinds of sources before loading DWH. Staging contains only current cleaned,p
rofiled data whereas DWH contains all historical cleaned,profiled data.
Main purpose is we cannot compare source format data with DWH directly as source
data will not be a proper cleaned data which may not match with cleaned data st
ored in DWH. So, first we clean all source data and load temporarly to staging
from where we comapre with DWH using SCDs.

Operational Data Store (ODS)
======================
Operational Data Store (ODS) is layer of repository that resides between a data
warehouse and the source system. The current, detailed data from disparate sourc
e systems is integrated, grouped subject areas wise in the ODS. The ODS is updat
ed more frequently than a data warehouse. The updates happen shortly after a sou
rce system data changes.

Difference between ODS and DWH
=========================

One of the key challenges in a data warehouse is to extract data from multiple n
on-integrated systems to a defined single space, all the while ensuring linkages
between data elements such as customer, transaction, account and other dimensio
nal data. By providing a foundation to integrated operational results, the ODS e
asily helps resolve this challenge and is often regarded as a common link betwee
n banks operational systems and the data warehouse. However, it is always advisab
le to have a clear distinction between the ODS and the DWH decision systems. Whi
le the DWH should store only historical data, the ODS needs to be more current i
n nature. It should be designed to rapidly update data so as to provide snapshot
s in an online, real-time environment.
There is also a marked trend towards keeping both the ODS and warehouse in the s
ame place as the interchangeability is very thin
Often, to avoid investing in technology to develop an ODS, banks end up using th
e old warehouse for operational reporting requirements as well. However, it need
s to be remembered that at the technology front, both the ODS and DWH need diffe
rent architectures. While the ODS needs update and record oriented functionality,
the warehouse requires a load and access oriented functionality.

An important recent trend is of keeping some functionalities of a DWH in the ODS
, such as summary or aggregated data for easy and fast query. Although it is ver
y difficult to have both current and summary data stored in an ODS, this helps t
o meet quickly and easily, different requirements from the same place. However,
as the ODS data gets regularly updated, it is important to realize that the accu
racy of the summary data in an
ODS is only for the specific point when it is accessed. So compared to a DWH, th
e reconstruction of same balances in an ODS will be difficult.
Essentially, the ODS derives some of its essence from that of a data warehouse w
ith certain key differences. An ODS can be defined as an application that enable
s any institution to meet its operational data analysis and processing needs. Th
e key features of an ODS are as below:
Subject oriented
Integrated
Volatile - updated data
Near time or current time data

Notably, it is the last two points that differentiate an ODS from a warehouse.
When we say an ODS needs to be subject oriented, it means that the basic design
should cater to specific business areas or domains for easy access and analysis
of data. It will be easier to store data in sub segments like customer, transact
ion and balance with an inter linkage between all the segments. That brings the
requirement of using a common key for all data files to establish the linkage. A
long with using the key, aggregation of data for certain time blocks as discusse
d above, it will also be helpful for easy and fast recovery of query results.

The following key functions are generally required for a successful ODS
Converting data
Selection of best data from multiple sources
Decoding/encoding data
Summarizing data
Recalculating data
Reformatting data and many more.

Apart from changes in the structure of an ODS in terms of handling large volume
volatile transactional data from multiple sources with a scope of aggregating th
em as well, there are also infrastructure related changes happening today. While
the basic technology of loading and processing high volume data remains the sam
e, when the system of records changes a little, the impact on the ODS can be ver
y significant and has to be managed very carefully. Therefore, even if the syste
m of records may comprise a very small portion of the system, it should be ensur
ed that the underlying infrastructure is able to handle high volumes at all time
s.
How to start building an ODS
It is advisable to start building a subject-oriented small ODS in the first plac
e with customer centric activities and transactions. Later, with integration in
mind, other subject areas can be added.

Conclusion
=======

Although there are many similarities between a data warehouse and an ODS and it
is often difficult to differentiate between them, there is a specific need for a
n ODS in banks today where large volume of data is required to be processed in a
fast and accurate manner to support the operational data reporting and analytic
al needs of the organization. As ODS contains specific data required for a set o
f business users, it helps in reducing the time to source information and with t
he added functionality of summary or aggregated data, it will definitely be an i
mportant element in any large banks IT systems infrastructure.
-----------------------------------
Please find the questions mentioned below
1)Slowly changing dimension types
2)DML statement
3)DDL statement
4)write a syntax to create database
5) differences between count(*) and count(column)
6) differences between union and union all
7) what is decode
8) what is NVL
9) Arrange the values in ascending order 10,5,20,15,null
10)Primary and surrogate key
11)what is group by clause does?
12)what is difference between truncate and delete
13)what is difference between OLAP and OLTP ?
14)Architecture of Datawarehouse
15)how do you test the data validation from source as (flat file and xml) to sta
ging tables.
16)what is shedule and time bound the mentioned above question asked in two roun
ds.all the best