Structured Data

This is the most common data classification we can find. Been here very long and it will be here to stay amid incoming new waves of technology in data processing. Reason is simple. It is reliable, mature, fast and scalable for transactional data. For more details about this data type, keep reading!

Introduction

We can easily identify this type by looking at Microsoft Excel. There are columns to identify the measurements and rows to store individual records. Each row can have one or more columns. Using Excel, we can easily add more columns until. Similarly, we only need to add more rows if we want to store more records.

An Excel Story

However, storing data in Excel is not robust. People can put wrong data in some columns. Of course we can put some restrictions in each column, but it will be tedious process. Even after we have this rules, the maximum row numbers and columns will have performance issue. You can try to have 1 million rows in Excel with about 10 to 20 columns. Try to open the file and calculate how much time it needs to load properly. Now imagine if you have some formula for some columns. It can crash quite easily.

Say Hi to Database Technology

This is where the database technology come in play. We need to put the restrictions from the beginning and then we can add as many rows as we want. Of course at some point you will need a bigger computing power and storage. But the data integrity will be there and it’s not easily tampered without using SQL operations. And the technology also has security group which can control who access what under Create, Read, Update and Delete policy.

Strength

This technology is really powerful for transactional operations. Take an example of banks. When we want to transfer some money to another account, the database will lock your account, then deduct the amount from it, lock the destination account, add this amount to this destination and commit the result. If and only if the whole process is successful, the locking mechanism will be released. If not, the whole process is back to the original state, meaning no money deducted or added for both account. And all of these can happen in less than 100 milliseconds (0.1 second), even in some cases in can be below 10 milliseconds.

Limitations

With the requirements to define the columns and data type restrictions at creation, this technology does not support very well for the dynamic nature of new data type emerged in 2000s. We started seeing pictures, videos and sound recordings are generated from the smartphone. This kind of data introduce a new challenge in storing them efficiently since we can not define how long should a video is or how many pixels should be in one image

See you in the next post!