qbr v1.0.0

R
javascript

Quick post to highligh some of the major updates in {qbr} v1.0.0

Published

2022/02/09

{qbr} v1.0.0

Recently I made some major changes to the qbr package I released back in June 2020. As a quick refresher, qbr is an R wrapper around the popular javascript library queryBuilder. In short, it provides a nice UI for constructing queries that can be used to filter data and also plays nicely with Shiny as its two main dependencies are bootstrap and jquery.

I initially developed qbr using the htmlwidgets framework, guided by an earlier attempt at integrating the library with R by @harveyl888. However after using qbr in a couple of projects at work, I found there were a few limitations, particularly when it came to updating the widget with new filters. So this latest attempt moves away from the htmlwidgets approach and instead uses a custom input binding for Shiny, and implements some additional functionality, all with the aim of making the library easier to use and customise as required.

So, whats new?

Below are just a few of the highlights in the new version of qbr - checkout the documentation for more details!

Support for different bootstrap versions

While not “officially” supported by the original js library, some kind folk have made suggested some fixes that enable the library to be used with bootstrap 4+. Within the new useQueryBuilder() function you can set bs_version to one of “3”, “4” or “5” (default is “3”) in order to make sure the builder looks and functions ok with newer versions of boostrap. This is especially useful if you’re using bslib to update the bootstrap version used by shiny!

Easily update the builder with updateQueryBuilder()

The original library has many methods that can used to on an existing builder. Currently updateQueryBuilder() supports the following,

  • reset: Removes all rules from the builder,
  • destroy: Self explanatory..!,
  • setRules: Load a predetermined set of rules into the builder.
  • setFilters: Remove all existing filters and add a new set of filters to the builder
  • addFilter: Add a filter to the existing filters

I’ve found these to be particularly useful when you need to builder to update depending on the data or a users actions in the app.

Choose your output

In the queryBuilderInput() function, you can use the return_value argument to choose the output that gets returned to shiny. This can be one of:

  • "r_rules": Returns the filter condition formatted for R
  • "sql_rules": Returns the filter condition formatted for SQL
  • "rules": list of conditions, default output from original queryBuilder library
  • "all": Return all of the above in a named list.

The aim here is to try and give the user more control over what output is returned.

Conditional dependencies for widgets

queryBuilder supports many different plugins and widgets that depend on additional dependencies. In the previous version of qbr, all of these dependencies were loaded in when you used the builder, even if you didn’t use any of the plugins or widgets! This led to a noticeable increase in loading time, and was just unnecessary. Now, only the dependencies for the requested plugins will be loaded, which should improve the loading time of the builder and is generally more efficient.

Demo

Below is a demo app using a couple of different configurations available.

Summary

So there it is! I still have a few ideas that I’d like to implement and generally make the package more robust, but in the mean time I’d love any feedback on the package, so feel free to drop a comment or open an issue on github. Thanks for reading!