What are some methods for optimizing Oracle databases for large data inserts?

For large data inserts I can suggest you few things like:

Use trigger(PL/SQL)
Use APPEND hint
Remove indexes on tables

Firstly, while you use triggers in the table it could leave data to be logically corrupt. And it will then perform insert in a very conventional way. Which is a time consuming process and won’t helps us!

Secondly, using APPEND hint will help us to an extend.

So, APPEND hint tells the optimizer to perform a direct-insert into the table, which improves the performance.

Now there is a way which we could achieve this by minimizing the Redo generation.

What Redo do is; it basically ensures the recoverability of data to the database. It writes down every transaction to the archive log.

Let’s take a scenario, where if the database is running on the NOARCHIVELOG mode, using APPEND hint will reduce the redo generation i.e; it won’t write into the archive log anymore and thus increases the speed.

But then it won’t be able to recover at any point in time if your data is ambiguous.

Furthermore, in the ARCHIVELOG mode, using APPEND hint won’t help us and will not reduce the redo generation until and unless your table is in NOLOGGING mode.

But also, in NOLOGGING mode our data is no longer protected. In fact, in OLTP it is very rare to see your table to NOLOGGING mode for obvious reasons.

Lastly, yes we can drop the indexes, constraints when bulk inserting into tables. Coz writing down index for every row formed will depreciate the performance of our db.

So, Ya! These are the methods for optimizing Oracle database for large data inserts.

I hope you got this!

Cheers! 👍

Link to my space: Click here

Comments

Machine Learning model for predicting 'Salary' of an Employee based on 'YearsofExperience'

“ Data really powers everything that we do .” — Jeff Weiner In the 21st century, Data is one of the most valuable entity anyone can have! There is loads-and-loads of data generated everyday. And to process this huge amount of data we need people who have expertise in it, who by the way are called as Data Engineers. Data Engineer collects the raw data, process it for further use; but we need an Analytic process which will automatically predict the data based on the previous one. And here's how 'Machine Learning' comes into the picture. "Machine Learning allows us to make highly accurate predictions based on the Historical Dataset which is used to train the machine learning model." Today let us look at a similar ML model to predict the 'Salary' of Employees based on 'YearsofExperience'. (P.S: I've provided pdf link at the very bottom of this page for clear understanding) 1) import the required modules 2) read the csv file 3) plot the graph 4) us...

Introduction to SQL

Q.) What is SQL ? Ans.) SQL(Structured Query Language) is a standardized language to communicate with the database. With the help of it we can retrieve data from the database. SQL not only allows us to read the data but also allows us to write the data in the database. Data in the database is stored in the form of tables. We can Select, Insert, Update, Delete, Create, Alter, Drop and can perform many more operations on the table. And these are called as SQL Statements. * SQL Statements are classified mainly into 4 categories :- 1) DML (Data Manipulation Language) 2) DDL (Data Definition Language) 3) DCL (Data Control Language) 4) TCL (Transaction Control Language) * Under DML we have :-(No Autocommit) 1. Select 2. Insert 3. Update 4. Delete 5. Merge * Under DDL :-(Allows Autocommit) 1. Create 2. Alter 3. Drop 4. Truncate 5. Flashback * Under DCL :-(Allows Autocommit) 1. Grant 2. Revoke * Under T...

Slicing in Python

* Let's first understand how a slicing is been done-- Consider a string namely==>> string = 'python_developer!' If we print this then-- # Output: python_developer! >> Now lets do some slicing here-- print (string[0:17:2]) >> Then the-- # Output: pto_eeoe! >> Lets see how it works-- So basically the print statement is in the form of print(string[a:b:c]) >> Now here -- a: Starting position b: Ending position c: Steps taken >> If we take out length of the given string then-- print (len(string)) # Output: 17 * The total length of the given string is 17. Therefore in our given problem the string will be printed from 0 index to 16 index and would take 2 steps. >> Hence the output-- # Output: pto_eeoe! >> Now the print statement that we have used here is-- print (string[0:17:2]) >> We can also use-- print (string[:17:2]) >> It will give the same output-- ...

code feat

Search This Blog