Web Application Generator

Apps created with Code On Time generator will automatically compose up to nine charts based on the columns present in your tables. If the default charts are insufficient, then the developer may choose to define their own charts for each data view.

Charts are configured using a simple tagging language applied on the data fields of a view. Let’s start configuring charts for the Orders page of the Northwind sample app.

Creating a Chart

Start the Project Designer. In the Project Explorer on the right side of the screen, switch to the Controllers tab and double-click on the Orders / Views / grid1 / CustomerID data field.

Change the following property:

The tag does the following: “pivot1” will declare that the data field CustomerID will be used for pivot with ID of “1”. The keyword “row1” declares that the data field will be used to define the first row. Finally, the keyboard “column” declares the chart type.

Press OK to save the data field. On the toolbar, press Browse to regenerate the app and open it in the default browser. When the page comes up, navigate to the Orders page. There will be a single chart of type “column” available.

The data behind the chart is shown in the table below:

The first column shows the values of each row (the customer company name). The second column is the actual value used to render the chart. No value field was specified by the developer, so a count of records has been used.

Sorting

Notice that the order of rows is alphabetical. Let’s sort the rows in descending order by the value in the second column. Replace the tag using the method described above with the following:

The image below shows the new chart, with the rows sorted.

The data will look like the following:

This has given us the correct order, but this chart can still be improved. Notice that there are too many columns displayed, making it hard to read. Let’s reduce that number.

Reducing the Result

Replace the tag with the following:

The “topX” keyword will only take the top X number of rows, and will throw away the rest. The new chart can be seen below.

The table can be seen below.

There are now ten rows of data available, as expected.

It may be possible that the user does not want the rest of the rows to be hidden, but grouped into one last column.

The “Other” Column

Using the “other” keyword in combination with “topX” will only display the top X number of rows. For rows that do not make the cut, their values will be summed up into an “Other” row. Use the following tag:

The new chart will display this “Other” column.

The values in this “Other” column will also be present in the chart data.

Defining Titles

Notice that a title has been automatically generated from the chart definition. The title, as well as the horizontal and vertical axis labels, can also be defined.

Replace the tag with the following. Notice that multiple space-separated tags are defined.

The new chart will use the specified titles:

The new title will be shown above the table as well:

This chart looks complete. Let’s add another chart that shows how many orders were received over time.

Working With Dates

Orders table in the Northwind database contains the OrderDate field. Let’s show how many orders were received per quarter. Apply the following tags to the relevant data fields:

The “quarter” keyword will group the values by quarters. The “line” keyword is the chart type.

In this chart, the orders have now been grouped by the quarter. Notice that there are no years – we need to define another group in order to split the orders by year, and then by quarter.

Make the following changes:

This configuration will define a chart that uses the OrderDate field twice – first time to group by year, and then to group by quarter. The results can be seen below.

The new chart now shows quarters from each year that Northwind has been in business.

Revealing Gaps

Suppose that the data has some gaps in the time period. The example below was filtered to customer “Frankenversand”, and it is clear that “1997, Q1” is missing. A missing row is easy to spot in the table, but it would be harder to spot in the chart.

Use the “all” keyword to ensure that gaps are displayed in the result.

The chart now presents gaps in the data.

This chart looks good with the current data set. What happens when the data takes on a different shape?

Automatic Date Sizing

The view has been filtered to only include data for 3 months. The end result is not so useful or attractive when using year/quarter.

The table is equally unhelpful.

This is where the “date” keyword comes to the rescue. Instead of manually specifying the date groupings, we will command the API to compose several result and use the best fit. Replace the tag with the following:

It appears that the chart using year/month/week has the most useful presentation according to the algorithm.

The table now reveals more useful information as well.

Specifying a Column Field

It may be helpful to split apart each data point by introducing a column field. Let’s show the number of orders from the top three customers. Change the tags for pivot2 as follows:

The chart will now show a breakdown of orders from the top three customers at each date.

The table will now have multiple columns.

Specifying a Value Field

Instead of simply showing a count of orders, let’s use the sum of Freight as the value. To do this, we will need to define the data field Freight as a value field. Specify the tags below. The “sum” keyword determines the type of value. By default, the chart will use “count”.

The chart will now display values according to the sum of Freight field.

The table will reveal the correct values.

Data views in an app created with Code On Time may present users with the view style called “Charts”. This view style is either enabled automatically by application at runtime or by developers at design time. The purpose of the view style is to provide instant insight into data.

If the definitions of charts are not a part of application design, then the application framework will compose up to nine charts based on the properties of fields in the data view if possible.  The app will attempt to define charts after examining the presence of lookup, date, or numeric fields in the view. First, it will try creating charts with dates as the rows, and lookups as the values. Then, it will pair a few numeric fields with a lookup.

The picture shown below displays the charts automatically created for the Orders controller of the Northwind sample application.

The automatically created definitions for all of the nine charts are shown below. Each chart is described with one or two tags that start with “pivot-“ keyword followed by properties.  The elegant simplicity of the tag language makes possible sophisticated collections of charts defined both by developers and end users.

Application framework performs efficient server-side data pivots following the tag specification. Data pivots for all charts are performed on the server simultaneously. Data pivots take into account application-level and user-defined filters.

Let’s see how each chart works.

Chart 1 – Stacked Column + Dates

This chart uses the first date field, OrderDate, as the rows. The “pivot1” property assigns the field to the pivot with ID of “1”. The “row” property sets the field as a row in the pivot, and assigns it index “1”.

The property “date” instructs the application to try several date bucket groups and choose the best one available based on the number of rows in the output. Application framework tries not to have too many or too few “date” rows in a pivot. For the dataset in the picture, the framework has produced the pivoted data as if tags “pivot1-row1-year pivot1-row2-month” were specified.

The “all” property ensures that any empty gaps in the data will be included in the pivot dimension.

The first lookup field, CustomerID, is used for the columns. The columns are sorted in descending order by the value of each column. Then, only the top five customers are kept, and the rest of them are dropped from the output. The “columnstacked” property defines the type of chart.

When a value field is not defined for the chart, a count of the first available field is used.

Chart 2 – Area + Dates

This chart uses the second date field, RequiredDate, as the rows.

The second lookup field, EmployeeID, is used for the columns, and the top seven employees sorted in descending order are used. The “area” property specifies the chart type.

Chart 3 – Column + Dates

This chart uses the third date field, ShippedDate, as the rows of the output.

The ShipVia lookup field is used, and only the top five values are kept. The type of the chart is defined as “column”.

Chart 4 – 3d Pie

The fourth chart uses CustomerID as the row values. The top ten values are displayed. The “other” property commands the server to sum up the rest of the values in an “Other” row. The type of chart is “pie3d”.

When no column field is defined, there will be only one column to display the value.

Chart 5 – Columns

The fifth chart uses EmployeeID lookup field for the rows. The top ten values are displayed, and the rest are summed up into an “Other” row. The type of chart is “column”.

Chart 6 – Line

The next chart uses the OrderDate field as the rows. The “date” property will generate multiple sets of data, differentiated by bucket size, and use the best result for the chart. The property “all” will ensure that there are no missing values in the date range. The type of chart is “line”.

Chart 7 – Columns

This chart uses the RequiredDate field for the row values. The type of chart is “column”. The server will compose multiple results and pick the best one that fits in the graph when “date” is used. The chart will not miss any gaps in dates with the “all” keyword.

Chart 8 - Area

This chart shows values grouped into rows by the ShippedDate field.

Chart 9- Donut

This chart uses the ShipVia lookup field for the rows of the result. The “top10” tag is disregarded as there are less than ten values in the result.