Learn Inverted Index| Building a search index from scratch -2#computerscience #python #searchengine

Añadido: 2026-06-02

[00:00:00]Welcome to the second video of this playlist where we build our own search engine from scratch. So, in this video we are going to implement a method which is known as inverted index. So, as you have seen in the first video, what we did was we were going through all the documents. So, basically if I have document 1 2 3, we were going through all the documents to search for all the tokens which we have in our search query.

[00:00:18]That is very not very much in optimal because we need to search the whole document for every term. So, it's basically if you have X terms and you have approximately if you have let me just write it over here in the readme. So, if you have X terms and you have Y documents and if each of the documents have like Z words, unique words, or maybe just words actually. So, total the total time complexity will be parallel complexity will be big O of X into Y into Z. So, this is really in optimal as you can see because it takes three parameters and if we have like a huge number of documents or maybe a lot of words in each documents for every search for every search query. It has to go through this whole big time complexity to give you the result.

[00:01:05]So, with inverted index, so what is an inverted index? Basically, instead of having the document ID as the key and the words as the value, what we have here is that the terms individually the individual tokens. For example, this is cat, dog, and fish. These are the keys and in the values we have the document IDs like this term is present in these specific IDs. So, if you build the index once, we can reuse it again and again and again. So, how this will help is that for every search we can directly go For example, if I have the query which has X words, so for every word I can directly look up the map which is there of this specific word. So, there are X terms, Y documents. So, with inverted index time complexity it is simply big O of X into the number of unique words which are there in documents, Z. So,