Lesson 9

<- Back to previous lesson

The goal of this lesson is to enable search within MongoDB Atlas by creating a search index. We should be able to search for flutters by taking advantage of fuzzy searching.

Be sure to switch to the 9-lesson branch in your local environment.

1. From the Atlas dashboard, select Database from the left menu.
2. Click the Browse Collections button on your cluster.
3. Select the Search tab.
4. Click the Create Index button.
5. Choose the Visual Editor and click Next.
6. You can name your index or leave it set to default.
   • If you alter the name, take note of the new name for later.
7. Choose the social_butterfly database and flutters collection, then press Next.
8. Press Create Search Index.

Use the one of the other endpoint files as a starter to create pages/api/flutter/[term].js.

This route will receive the search term as a query parameter.

This route should get all flutters that match the search term.

The term can be retrieved from the query parameter: req.query.term.

The Data API endpoint for this route should be /aggregate.

To use the search index, include a pipeline in the request body.

Here is an example pipeline using a search index:

[
  {
    $search: {
      index: "default", // The name of the index
      text: {
        query: "Hello", // The search term
        path: {
          wildcard: "*", // The fields to search
        },
      },
    },
  },
];

After the search pipeline stage, add a sort stage that sorts descending on the postedAt field.

Test your search functionality by running the application and using the search input in the header.

Notice that the search results are very strict. You have to be exact with the search term.

We can fix this by adding a fuzzy parameter to our search pipeline stage.

case "GET":
  const term = req.query.term;
  const readData = await fetch(`${baseUrl}/aggregate`, {
    ...fetchOptions,
    body: JSON.stringify({
      ...fetchBody,
      pipeline: [
        {
          $search: {
            index: "default",
            text: {
              query: term,
              path: {
                wildcard: "*",
              },
              fuzzy: {}
            },
          },
        },
        { $sort: { postedAt: -1 } },
      ],
    }),
  });
  const readDataJson = await readData.json();
  res.status(200).json(readDataJson.documents);
  break;

Great job! Let's move on to the next lesson ->

Be sure to commit your branch changes and switch to the 10-lesson branch in your local environment.