Index size bigger than the orginal data

Is it normal to get an index file which is larger (in size) than the original data file?

I have JSON file with ~ 1MB of size (about 58,278 words). When Trying to build an index for it using the following code:

const index_ar = new FlexSearch({
  tokenize: "strict",
  rtl: true,
  split: /\s+/,
  doc: {
    id: "id",
    field: [
      'title',
      'incident_date_time',
      'location:name'
    ]
  }
});
 index_ar.add(data);

The index file size is ~ 2.1MB! I inspected the file size using the following method:

const exportedIndex = index_ar.export();
fs.writeFileSync('exported.json', JSON.stringify(exportedIndex));

Is there any wrong in the code? or it’s normal to get an index size bigger than original data?

[desjob]

I would say that it is expected that the index size is always larger as the original data, because the index will duplicate data to be able to search fast.

Think of it as adding a traditional index to an existing book: for each word that occurs in the book, you add a list on which pages it occurs. Now your book is several pages thicker!

[tareefdev]

This is a really nice option, thanks for the great work.

I want to emphasize again on that particular use case, when using very basic functionalities from FlexSearch; (“strict” tokenizer, non-contextual, split by words). In this case, I don’t see any difference (in performance and data size) between:

• Building an index and load it with a copy of my data
• Just loading the data and then using a framework (React) to filter and display it.

What benefits of using FlexSearch here?