I have seen so many e-commerce websites that provides search box to search products. In that search features most of the search fields are auto-complete. If we enter a letter on field, then it will show the data which is including that letter as suggestions from database. As I know basics on developing that functionality.
But what if database contains huge amount of data?
For example, e-commerce websites like flipkart and amazon had a lot of products in their database. so, if user enter a letter in search field, it have to search for data including that letter among all the data in database and match data including that letter and display data as suggestions. The websites are processing it within nano seconds of time. I wonder how they achieved that functionality? I can't understand what are the technologies they are using.
As a learner I wanna know the functional flow and if possible demo for that feature.
I think your question can be divided into two parts. 1) how to design the database for the search technology. 2) how to implement an effect search... They belong to the field of search engine technology.
About the Q1, you can create a table to save the keywords for search, and in the table, you'd better to design a column or similar method to describe the "search-weight". As well known, a view is a practical solution to accelerate the access of the data.
About the Q2, the search engine technique is No longer mysterious, some open source projects can simulate the feature of search engine, such as Apache Lucene, visit please Apache Lucene.
more discuss: And specially, in your front system for example, the ASP/JSP or even simple HTML page, you should use some scripts e.g. Ajax, to popup, drawdown, of caurse, simple DOM Javascript+DIV can reach it too, but with jQuery or other libarary can make it easily. Here is an example.
Here is the backend system example To reduce the burden on the host and reduce the requirement of network's bandwidth, the front javascript should active the autocomplete feature with more than three characters.
Please pay attention in your actual application, that your server has calculation's limitation, and the client page has usually many elements, all will reduce user friendliness. Please do not make the request and response too complex.
An alternative simulation can be: make a FIFO logic, save some usual search keyword in the "cache" or temp-table|view, and the amount of data will be reduced.
There are too many solutions, I can only think of these tricks at this moment. regards