Adding DashboardElements to a Dashboard via API

In the 5.14 release, we have added newly documented experimental endpoints for managing Dashboard Elements via the 3.1 API. You can learn more about the change described for the API Explorer

This code snippet is from test code for a Python SDK prototype (still rapidly and drastically changing, with no public or beta ETA) that shows the basics of how to add elements to a new or existing dashboard. Most of the work on the SDK at this point is just making calls to the swagger python codegen API more convenient, but this code does show the kinds of values you need to provide to the API, and a method for adding and manipulating dashboard elements. I hope it helps clear up some things for you.

def draw_dashboard(sdk, dashboard):

    top_center = sdk.create_dashboard_element(

        dashboard_id=dashboard.id,

        type="text",

        title="What's all this 1",

        title_text="This is the title of a text element",

        subtitle_text="Put any subtitle here",

        body_text="Put something else in the body"

    )

    # vis types from https://docs.looker.com/reference/dashboard-reference/lookml-visualization-reference#type-for-elements

    # text | table | single_value | looker_single_record | looker_column | looker_bar |

    # looker_scatter | looker_line | looker_area | looker_pie | looker_donut_multiples |

    # looker_geo_coordinates | looker_geo_choropleth

    brand_query = sdk.create_query(

        model="thelook",

        view="products",

        fields=["products.brand", "products.count"],

        vis_config={

            'type': 'looker_column'

        }

    )

    # vis_config={

    #     'stacking': '',

    #     'show_value_labels': False,

    #     'label_density': 25,

    #     'legend_position': 'center',

    #     'x_axis_gridlines': False,

    #     'y_axis_gridlines': True,

    #     'show_view_names': True,

    #     'limit_displayed_rows': False,

    #     'y_axis_combined': False,

    #     'show_y_axis_labels': True,

    #     'show_y_axis_ticks': True,

    #     'y_axis_tick_density': 'default',

    #     'y_axis_tick_density_custom': 5,

    #     'show_x_axis_label': True,

    #     'show_x_axis_ticks': True,

    #     'x_axis_scale': 'auto',

    #     'y_axis_scale_mode': 'linear',

    #     'ordering': 'none',

    #     'show_null_labels': False,

    #     'show_totals_labels': False,

    #     'show_silhouette': False,

    #     'totals_color': '#808080',

    #     'type': 'looker_column',

    #     'show_row_numbers': True,

    #     'truncate_column_names': False,

    #     'hide_totals': False,

    #     'hide_row_totals': False,

    #     'table_theme': 'editable',

    #     'enable_conditional_formatting': False,

    #     'conditional_formatting_include_totals': False,

    #     'conditional_formatting_include_nulls': False,

    #     'hidden_fields': ['users.id'],

    #     'series_types': {},

    #     'y_axis_orientation': ['right', 'left'],

    #     'y_axis_max': ['10', '100']

    # }

    bottom_left = sdk.create_dashboard_element(

        dashboard_id=dashboard.id,

        type="vis",

        title_text="Brand Count",

        title="Brand Count Title",

        subtitle_text="Vis Subtitle",

        body_text="Vis Body",

        query_id=brand_query.id

    )

    bottom_right = sdk.create_dashboard_element(

        dashboard_id=dashboard.id,

        type="data",

        title_text="Users Data",

        title="Users Data Title",

        subtitle_text="Data Subtitle",

        body_text="Data Body",

        query_id=query.id

    )

    dashboard = sdk.dashboard(dashboard.id)  # fetch updated dashboard

    layout = dashboard.dashboard_layouts[0]

    top_height = 2

    for comp in layout.dashboard_layout_components:

        if comp.dashboard_element_id == top_center.id:

            sdk.update_dashboard_layout_component(

                comp.id,

                dashboard_layout_id=layout.id,

                dashboard_element_id=top_center.id,

                row=0,

                column=0,

                width=24,

                height=top_height

            )

        elif comp.dashboard_element_id == bottom_left.id:

            sdk.update_dashboard_layout_component(

                comp.id,

                dashboard_layout_id=layout.id,

                dashboard_element_id=bottom_left.id,

                row=top_height,

                column=0,

                width=12,

                height=8

            )

        elif comp.dashboard_element_id == bottom_right.id:

            sdk.update_dashboard_layout_component(

                comp.id,

                dashboard_layout_id=layout.id,

                dashboard_element_id=bottom_right.id,

                row=top_height,

                column=12,

                width=12,

                height=8

            )

    return sdk.dashboard(dashboard.id)  # fetch updated dashboard



def create_filters(sdk: LookerSDK, dashboard: Dashboard):

    for f in dashboard.dashboard_filters:

        sdk.delete_dashboard_filter(f.id)



    dashboard_filter = sdk.create_dashboard_filter(

        dashboard_id=dashboard.id,

        name='Brand match',

        title='Brand match',

        row=0,

        type='field_filter',

        model='thelook',

        explore='products',

        dimension='products.brand',

        allow_multiple_values=True,

        default_value='A%,B%,C%,D%',

    )



    # activate the filter for the brand tile

    [brand_tile] = [tile for tile in dashboard.dashboard_elements if tile.title_text == 'Brand Count']

    sdk.update_dashboard_element(

        brand_tile.id,

        listen={'Brand match': 'products.brand'}

    )

    return sdk.dashboard(dashboard.id)  # fetch updated dashboard

Hello! Still having trouble with this. How can I just add a look to a dashboard. I am using the REST API to do this. It seems like it should be as simple as passing the dashboard_id with the look creation request or sending a PATCH request to update the dashboard with a look_id.

The documentation needs better example queries. Specifically with things like how to set a visualization type. What those visualization types are… etc. All you get with the interactive API explorer is the response model. That is not sufficient for knowing how to send requests. It also requires developers to search all over the other types of documentation for hints of what to send and how to send it.

So I had a discussion with a very helpful Looker Representative… He give me some advice that may help future readers:

The model in the API Explorer has required fields denoted. You can use this to build a basic request and once you have that the API gives much more helpful error messages often times listing possible options for certain fields if you don’t provide a proper one.

To answer the OP’s question though these is an endpoint for adding looks to dashboards
e.g.
POST /dashboard_elements

{
    "look_id": 100,
    "dashboard_id":222,
    "type":"vis"
}