Autocomplete

Index

Using fixed collection or collectionAsync
Editor Options (AutocompleterOption interface)
Using Remote API
Force User Input
How to change drop container dimensions?
Animated Gif Demo
- See the Editors - Wiki for more general info about Editors (validators, event handlers, ...)

Demo

Introduction

AutoComplete is a functionality that let the user start typing characters and the autocomplete will try to give suggestions according to the characters entered. The collection can be a fixed JSON files (collection of strings or objects) or can also be an external remote resource to an external API. For a demo of what that could look like, take a look at the animated gif demo below.

We use an external lib named Autocomplete (aka autocompleter on npm) by Kraaden.

Using `collection` or `collectionAsync`

If you want to pass the entire list to the AutoComplete (like a JSON file or a Web API call), you can do so using the collection or the collectionAsync (the latter will load it asynchronously). You can also see that the Editor and Filter have almost the exact same configuration (apart from the model that is obviously different).

Component

interface Props {}
interface State {
  columnDefinitions: Column[];
  gridOptions: GridOption;
  dataset: any[];
}
export class GridBasicComponent extends React.Component<Props, State> {
  componentDidMount() {
    this.defineGrid();
  }

  defineGrid(): void {
    const columnDefinitions = [
      {
        id: 'countryOfOrigin', name: 'Country of Origin', field: 'countryOfOrigin',
        formatter: Formatters.complexObject,
        dataKey: 'code', // our list of objects has the structure { code: 'CA', name: 'Canada' }, since we want to use the code`, we will set the dataKey to "code"
        labelKey: 'name', // while the displayed value is "name"
        type: FieldType.object,
        sorter: Sorters.objectString, // since we have set dataKey to "code" our output type will be a string, and so we can use this objectString, this sorter always requires the dataKey
        filterable: true,
        sortable: true,
        minWidth: 100,
        editor: {
          model: Editors.autocompleter,
          customStructure: { label: 'name', value: 'code' },
          collectionAsync: fetch('assets/data/countries.json'), // this demo will load the JSON file asynchronously
        },
        filter: {
          model: Filters.autocompleter,
          customStructure: { label: 'name', value: 'code' },
          collectionAsync: fetch('assets/data/countries.json'),
        }
      }
    ];

    const gridOptions = {/*...*/};

    this.setState((state: State) => ({
      ...state,
      gridOptions,
      columnDefinitions,
      dataset: this.getData(),
    }));
  }
}

Collection Label Render HTML

By default HTML is not rendered and the label will simply show HTML as text. But in some cases you might want to render it, you can do so by enabling the enableRenderHtml flag.

NOTE: this is currently only used by the Editors that have a collection which are the MultipleSelect & SingleSelect Editors.

this.columnDefinitions = [
  {
    id: 'effort-driven', name: 'Effort Driven', field: 'effortDriven',
    formatter: Formatters.checkmarkMaterial,
    type: FieldType.boolean,
    editor: {
      model: Editors.autocompleter,
      placeholder: '&#128269; search city',
      type: FieldType.string,

      // example with a fixed Collection (or collectionAsync)
      editorOptions: {
        showOnFocus: true, // display the list on focus of the autocomplete (without the need to type anything)
      } as AutocompleterOption,
      enableRenderHtml: true, // this flag only works with a fixed Collection
      // collectionAsync: this.http.get(URL_COUNTRIES_COLLECTION),
      collection: [
        { value: '', label: '' },
        { value: true, label: 'True', labelPrefix: `<i class="mdi mdi-plus"></i> ` },
        { value: false, label: 'False', labelPrefix: `<i class="mdi mdi-minus"></i> ` }
      ],
    }
  }
];

Editor Options (`AutocompleterOption` interface)

All the available options that can be provided as editorOptions to your column definitions can be found under this AutocompleterOption interface and you should cast your editorOptions to that interface to make sure that you use only valid options of the autocomplete library.

editor: {
  model: Editors.autocompleter,
  editorOptions: {
    minLength: 3,
  } as AutocompleterOption
}

Grid Option `defaultEditorOptions

You could also define certain options as a global level (for the entire grid or even all grids) by taking advantage of the defaultEditorOptions Grid Option. Note that they are set via the editor type as a key name (autocompleter, date, ...) and then the content is the same as editorOptions (also note that each key is already typed with the correct editor option interface), for example

this.gridOptions = {
  defaultEditorOptions: {
    autocompleter: { debounceWaitMs: 150 }, // typed as AutocompleterOption
  }
}

Using External Remote API

You could also use external 3rd party Web API (can be JSONP query or regular JSON). This will make a much shorter result since it will only return a small subset of what will be displayed in the AutoComplete Editor or Filter. For example, we could use GeoBytes which provide a JSONP Query API for the cities of the world, you can imagine the entire list of cities would be way too big to download locally, so this is why we use such API.

Remote API (basic)

The basic functionality will use built-in 3rd party lib styling that is to display a label/value pair item result.

Component

export class GridBasicComponent extends React.Component<Props, State> {
  columnDefinitions: Column[];
  gridOptions: GridOption;
  dataset: any[];

  defineGrid() {
    const columnDefinitions = [
      {
        id: 'product', name: 'Product', field: 'product',
        filterable: true,
        minWidth: 100,
        editor: {
          model: Editors.autocompleter,
          alwaysSaveOnEnterKey: true,
          editorOptions: {
            showOnFocus: true,
            minLength: 1,
            fetch: (searchText, updateCallback) => {
              // assuming your API call returns a label/value pair
              yourAsyncApiCall(searchText) // typically you'll want to return no more than 10 results
                 .then(result => updateCallback((results.length > 0) ? results : [{ label: 'No match found.', value: '' }]))
                 .catch(error => console.log('Error:', error));
            },
          } as AutocompleterOption,
        },
      }
    ];
    const gridOptions = {/*...*/};

    this.setState((state: State) => ({
      ...state,
      gridOptions,
      columnDefinitions,
      dataset: this.getData(),
    }));
  }
}

Remote API (basic with object result)

This is the preferred way of dealing with the AutoComplete, the main reason is because the AutoComplete uses an <input/> and that means we can only keep 1 value and if we do then we lose the text label and so using an Object Result makes more sense. Note however that you'll need a bit more code that is because we'll use the FieldType.Object and so we need to provide a custom SortComparer and also a custom Formatters and for them to work we also need to provide a dataKey (the value) and a labelKey (text label) as shown below.

this.columnDefinitions = [
  {
    id: 'product', name: 'Product', field: 'product',
    dataKey: 'id',
    labelKey: 'name', // (id/name) pair to override default (value/label) pair
    editor: {
      model: Editors.autocompleter,
      alwaysSaveOnEnterKey: true,
      type: 'object',
      sortComparer: SortComparers.objectString,
      editorOptions: {
        showOnFocus: true,
        minLength: 1,
        fetch: (searchText, updateCallback) => {
          // assuming your API call returns a label/value pair
          yourAsyncApiCall(searchText) // typically you'll want to return no more than 10 results
            .then(result => updateCallback((results.length > 0) ? results : [{ label: 'No match found.', value: '' }]); })
            .catch(error => console.log('Error:', error);
      },
    } as AutocompleterOption,
  }
];

Remote API with `renderItem` + custom layout (`twoRows` or `fourCorners`)

See animated gif (twoRows or fourCorners)

The lib comes with 2 built-in custom layouts, these 2 layouts also have SASS variables if anyone wants to style it differently. When using the renderItem, it will require the user to provide a layout (2 possible options twoRows or fourCorners) and also a templateCallback that will be executed when rendering the AutoComplete Search List Item. For example:

Component

export class GridBasicComponent extends React.Component<Props, State> {
  defineGrid() {
      // your columns definition
    const columnDefinitions = [
      {
        id: 'product', name: 'Product', field: 'product',
        filterable: true,
        minWidth: 100,
        editor: {
          model: Editors.autocompleter,
          alwaysSaveOnEnterKey: true,
          customStructure: {
            label: 'itemName',
            value: 'id'
          },
          editorOptions: {
            showOnFocus: true,
            minLength: 1,
            fetch: (searchText, updateCallback) => {
              yourAsyncApiCall(searchText) // typically you'll want to return no more than 10 results
                 .then(result => updateCallback((results.length > 0) ? results : [{ label: 'No match found.', value: '' }]))
                 .catch(error => console.log('Error:', error);
            },
            renderItem: {
              layout: 'twoRows',
              templateCallback: (item: any) => `<div class="autocomplete-container-list">
                <div class="autocomplete-left">
                  <span class="mdi ${item.icon} mdi-26px"></span>
                </div>
                <div>
                  <span class="autocomplete-top-left">
                    <span class="mdi ${item.itemTypeName === 'I' ? 'mdi-information-outline' : 'mdi-content-copy'} mdi-14px"></span>
                    ${item.itemName}
                  </span>
                <div>
              </div>
              <div>
                <div class="autocomplete-bottom-left">${item.itemNameTranslated}</div>
              </div>`,
            },
          } as AutocompleteOption,
        },
      }
    ];
    const gridOptions = {/*...*/};

    this.setState((state: State) => ({
      ...state,
      gridOptions,
      columnDefinitions,
      dataset: this.getData(),
    }));
  }
}

Remote API `renderItem` callback + custom layout (`twoRows` or `fourCorners`)

See animated gif (twoRows or fourCorners)

The previous example can also be written using the renderItem callback and adding classes, this is actually what Slickgrid-Universal does internally, you can do it yourself if you wish to have more control on the render callback result.

Component

export class GridBasicComponent extends React.Component<Props, State> {
  defineGrid() {
    const columnDefinitions = [
      {
        id: 'product', name: 'Product', field: 'product',
        filterable: true,
        minWidth: 100,
        editor: {
          model: Editors.autocompleter,
          alwaysSaveOnEnterKey: true,
          customStructure: {
            label: 'itemName',
            value: 'id'
          },
          editorOptions: {
            showOnFocus: true,
            minLength: 1,
            classes: {
              // choose a custom style layout
              // 'ui-autocomplete': 'autocomplete-custom-two-rows',
              'ui-autocomplete': 'autocomplete-custom-four-corners',
            },
            fetch: (searchText, updateCallback) => {
              yourAsyncApiCall(searchText) // typically you'll want to return no more than 10 results
                 .then(result => updateCallback((results.length > 0) ? results : [{ label: 'No match found.', value: '' }]); })
                 .catch(error => console.log('Error:', error);
            },
            renderItem: {
              layout: 'twoRows',
              templateCallback: (item: any) => `<div class="autocomplete-container-list">
                <div class="autocomplete-left">
                  <span class="mdi ${item.icon} mdi-26px"></span>
                </div>
                <div>
                  <span class="autocomplete-top-left">
                    <span class="mdi ${item.itemTypeName === 'I' ? 'mdi-information-outline' : 'mdi-content-copy'} mdi-14px"></span>
                    ${item.itemName}
                  </span>
                <div>
              </div>
              <div>
                <div class="autocomplete-bottom-left">${item.itemNameTranslated}</div>
              </div>`,
            },
          } as AutocompleteOption,
          callbacks: {
             _renderItem: (ul: HTMLElement, item: any) => {
                const template = `<div class="autocomplete-container-list">
                      <div class="autocomplete-left">
                        <!--<img src="http://i.stack.imgur.com/pC1Tv.jpg" width="50" />-->
                        <span class="mdi ${item.icon} mdi-26px"></span>
                      </div>
                      <div>
                        <span class="autocomplete-top-left">
                          <span class="mdi ${item.itemTypeName === 'I' ? 'mdi-information-outline' : 'mdi-content-copy'} mdi-14px"></span>
                          ${item.itemName}
                        </span>
                      <div>
                    </div>
                    <div>
                      <div class="autocomplete-bottom-left">${item.itemNameTranslated}</div>
                    </div>`;

                return $('<li></li>')
                  .data('item.autocomplete', item)
                  .append(template)
                  .appendTo(ul);
              }
          },
        },
      }
    ];
    const gridOptions = {/*...*/};
}

with JSONP

Example from an external remote API (geobytes) returning a JSONP response.

Component

export class GridBasicComponent extends React.Component<Props, State> {
  defineGrid(): void {
    const columnDefinitions = [
      {
        id: 'cityOfOrigin', name: 'City of Origin', field: 'cityOfOrigin',
        filterable: true,
        minWidth: 100,
        editor: {
          model: Editors.autocompleter,
          placeholder: 'search city', //  you can provide an optional placeholder to help your users
          editorOptions: {
            minLength: 3, // minimum count of character that the user needs to type before it queries to the remote
            fetch: (searchText, updateCallback) => {
              // assuming your API call returns a label/value pair
              yourAsyncApiCall(searchText) // typically you'll want to return no more than 10 results
                 .then(result => updateCallback((results.length > 0) ? results : [{ label: 'No match found.', value: '' }]))
                 .catch(error => console.log('Error:', error));
            }
          },
        },
        filter: {
          model: Filters.autocompleter,
          // placeholder: '&#128269; search city', // &#128269; is a search icon, this provide an option placeholder
          filterOptions: {
            minLength: 3, // minimum count of character that the user needs to type before it queries to the remote
            fetch: (searchText, updateCallback) => {
              // assuming your API call returns a label/value pair
              yourAsyncApiCall(searchText) // typically you'll want to return no more than 10 results
                 .then(result => updateCallback((results.length > 0) ? results : [{ label: 'No match found.', value: '' }]))
                 .catch(error => console.log('Error:', error));
            }
          },
        }
      }
    ];
    const gridOptions = {/*...*/};
  }
}

Autocomplete - force user input

If you want to add the autocomplete functionality but want the user to be able to input a new option, then follow the example below:

this.columnDefinitions = [{
  id: 'area',
  name: 'Area',
  field: 'area',
  type: FieldType.string,
  editor: {
    model: Editors.autocompleter,
    editorOptions: {
      minLength: 0,
      forceUserInput: true,
      fetch: (searchText, updateCallback) => {
        updateCallback(this.areas); // add here the array
      },
    }
  }
}];

You can also use the minLength to limit the autocomplete text to 0 characters or more, the default number is 3.

How to change drop container dimensions?

You might want to change the dimensions of the drop container, this 3rd party library has a customize method to deal with such a thing. Slickgrid-Universal itself is removing the width using this method, you can however override this method to change the drop container dimensions

this.columnDefinitions = [{
    id: 'product', name: 'Product', field: 'product', filterable: true,
    editor: {
      model: Editors.autocompleter,
      alwaysSaveOnEnterKey: true,

      // example with a Remote API call
      editorOptions: {
        minLength: 1,
        fetch: (searchTerm, callback) => {
          // ...
        },
        customize: (_input, _inputRect, container) => {
          // change drop container dimensions
          container.style.width = '250px';
          container.style.height = '325px';
        },
      } as AutocompleterOption,
    },
}];

Animated Gif Demo

Basic (default layout)

with `twoRows` custom layout (without optional left icon)

with `fourCorners` custom layout (with extra optional left icon)

PreviousEditors Nextold Date Picker (flatpickr)

Last updated 7 months ago

Demo

Introduction

Using collection or collectionAsync

Collection Label Render HTML

Editor Options (AutocompleterOption interface)

Using External Remote API

Remote API (basic)

Remote API (basic with object result)

Remote API with renderItem + custom layout (twoRows or fourCorners)

Remote API renderItem callback + custom layout (twoRows or fourCorners)

Autocomplete - force user input

How to change drop container dimensions?

Animated Gif Demo

Basic (default layout)

with twoRows custom layout (without optional left icon)

with fourCorners custom layout (with extra optional left icon)

Using `collection` or `collectionAsync`

Editor Options (`AutocompleterOption` interface)

Remote API with `renderItem` + custom layout (`twoRows` or `fourCorners`)

Remote API `renderItem` callback + custom layout (`twoRows` or `fourCorners`)

with `twoRows` custom layout (without optional left icon)

with `fourCorners` custom layout (with extra optional left icon)