This page explains how to configure a chart that uses a record type as its source.
Charts in Appian can be configured using one of two methods:
categories and series parameters in the chart.If the source of your chart is a record type that uses a database as the source, or a record type that has sync enabled, use the record type in the configuration. Once you select the record type, it's easy to reference record fields or related record fields in the chart's labels or calculations.
If the source of your chart is not a record type, or you need to perform advanced calculations or manipulations on the data prior to displaying it on your chart (e.g. calculate a running sum after querying the results), use the categories and series parameters to construct the chart using an expression. See examples for configuring charts using categories and series directly in the component pages for Bar, Column, Line, and Pie charts.
To easily configure your charts, Appian recommends using a record type as the source. To configure a chart to use record data, you will:
Appian makes it quick and easy to select a record type as the source of your chart. Simply choose Record Type as the source of your data and search for your record type. The selected record type determines which fields are available to display on the chart.
Once you select the source, there are two key configurations needed to display your record data: groupings and a measure. The combination of these configurations determine what data to query from the record type and how to display the results on the chart.
To configure a chart to use a record type as its source:
Depending on the chart you choose, you can configure additional chart properties like X-Axis and Y-Axis Titles, Reference Lines, Color Scheme, and more.
Once you select a source record type, record fields will auto-populate in design view for the primary grouping and the measure. Each of these parameters define a key aspect of the design for your chart.
Groupings allow you to determine how the data is grouped together in the chart. You can group by any field in the record type, including related record fields.
For example, say you have a record type of all open support cases and you want to see how many open cases are associated with each employee. You can use grouping to group open cases by employee.
Depending on the type of chart you use, the way you group your data and the complexity of your groupings will vary:
Line charts also support multiple groupings. Like bar and column charts, the Primary Grouping displays on the x-axis. We recommend using time series data for this grouping. A Secondary Grouping will show as multiple lines in a line chart.
When there are multiple measures in your bar, column, or line chart, you cannot add a Secondary Grouping.
To configure a chart's groupings in Expression Mode, see the Grouping Component or the examples below.
When configuring the grouping of a chart, you can also use a date interval to group dates together and view trends at a summarized level.
For example, suppose you would like to display a line chart with the number of cases created over time. However, the only relevant field you have is createdOn, which is a Date and Time field. If you chose to group by the createdOn field without an interval, it will only group together data that occurred at the exact same time. Instead of grouping by the full date and time, it makes sense instead to group at a higher level, for instance by showing cases created by month.
In this example, you can add an interval that will group the data by the month. There are several intervals available, and each provides a different perspective on your data. See the Grouping Component for a list of available interval values.
To choose an interval in Design Mode, click the edit icon next to your grouping. If you are grouping by a Date, Date and Time, or Time field, you can select an interval from the Time Interval dropdown.
Once you choose an interval, you can also define how the data is formatted using a Format Value. Some intervals have pre-defined formats that you can choose. If you select one of these formats, it will automatically localize the data based on language for the user viewing the chart. For example, if you select the long text version of date, the value in US English will show as April 13, 2021. However, a user with a French locale will see the same chart with the value 13 avril 2021.
You also have the flexibility to customize your date format. With the Use custom format option, you can use the following characters to format your date values:
| Unit of Time | Character | Valid Combinations |
|---|---|---|
| Year | y |
yy, yyyy |
| Month | M |
M, MM, MMM, MMMM, MMMMM |
| Day | d |
d, dd, ddd |
| Hour | h |
h, hh |
| Minute | m |
mm |
| Second | s |
ss |
| Day of Week | E |
EEE, EEEE |
| AM / PM | a |
a |
| Timezone | z |
z, zzzz |
You can combine these characters to create a variety of formats, and use special characters like / or - to separate your values. For example, d-m-yy will return 17-9-21, whereas dd mmmm yyyy will return 17 September 2021.
In addition to using Format Value to format your dates, you can also use this option to format text and integer fields used in your groupings.
For example, suppose the grouping field displayed on the chart is the status ID. The status labels exist in a constant, so the labels aren't available directly from the data source. Using the status ID as the grouping field would display on the chart like this:
To display the status labels instead, we can add an expression to the formatValue parameter of a!grouping():
1
2
3
4
5
6
7
8
9
a!grouping(
field: recordType!Cases.field.statusId,
alias: "status",
formatValue: index(
cons!MY_APP_STATUS_LABELS /* Status labels available in a constant*/
fv!value,
"Unknown Status"
)
)
Using the expression above, the chart will instead display like this:
A measure determines how to calculate the values of a field in a chart. A measure can calculate on any field in the record type, including related record fields, but it should usually be a numeric field.
For example, say you have a record type that lists all customer requests and you want to see how many requests have been marked as "Success", "Fail", or "Exception". You can use grouping to display requests by status and use measure to count the number of requests per status.
To configure a chart's measure:
If you use a bar, column, or line chart, you can display multiple measures if no Secondary Grouping field is provided. When multiple measures display on a chart, each measure displays in chronological order on the chart and legend, and is assigned a different color based on the color scheme.
To configure a chart's measure in Expression Mode, see the Measure Component or the examples below.
If your record type has data sync enabled, you also have the flexibility to add filters to your measures. By applying filters on a record field or a related record field, you can determine which values are included in the measure's calculation.
For example, let's say you want to create a column chart that shows the number of orders associated with each customer, and you also want to see how many of those orders have a priority level of "Critical".
In your chart, you'd use the Customer record type to group by the name field and use the relationship with the Order record type to create two measures. The first measure will calculate the count of order Ids. The second measure will also calculate the count of order Ids and add a filter on the label field from the Priority record type so you only count orders that have a priorityLabel = Critical.
Now, let's say you want to create another report that shows the number of orders associated with each customer, and how many customer orders include at least one out-of-stock item.
In this example, you'd use the Customer record type to group by the name field, and use the relationship with the Order record type to create two measures. The first measure will calculate the count of order Ids. The second measure will also calculate the count of order Ids and add a filter on the status field from the Order Item record type so you only count orders that have a status = Out-of-stock.
In this example, the second measure counts all orders with at least one item marked as "Out-of-stock". This is because there is a one-to-many relationship between the Order record type and the Order Item record type.
When you filter on a field from the related record type (the "many" side of the relationship), the filter returns all records from the base record type (the "one" side of the relationship) that have at least one related record that meets the filter condition.
See Aggregating on related record fields for more examples using related record fields in your filtered aggregations. See Filters to learn more about configuring filters in Basic or Expression mode.
A link in a chart can be used to drill down into a grid, or filter other components on an interface. Chart links consist of a dynamic link and the variable fv!selection that contains the current selected values.
The fv!selection variable returns a map with all information about the current selected item, which allows you to dynamically pass chart data. To save the current selected values, you'll create a local variable to store the values.
For example, suppose you have the following stacked column chart that displays a count of restaurants by region and type:
In Expression Mode, you'll see that the chart's link parameter is pre-populated with the a!dynamicLink() function, and the value parameter contains fv!selection.
To dynamically save the current selected values, you'll add a local variable, like local!selectedItem, and use it in the saveInto parameter:
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
a!localVariables(
local!selectedItem,
{
a!columnChartField(
data: recordType!Restaurants,
config: a!columnChartConfig(
primaryGrouping: a!grouping(
field: recordType!Restaurants.fields.region,
alias: "region"
),
secondaryGrouping: a!grouping(
field: recordType!Restaurants.fields.type,
alias: "type"
),
measures: a!measure(
field: recordType!Restaurants.fields.id,
function: "COUNT",
alias: "id_count"
),
link: a!dynamicLink(
value: fv!selection,
saveInto: local!selectedItem
)
)
)
}
)
Now, when you select the bar chart shown above, the following data is saved to the local variable:
Notice that the data that is saved depends on the alias defined for each field. If you add a drilling link, the alias field is required. To add an alias, click the edit icon next to Measure and enter a value in the Alias field.
Additionally, the saved value is always the unformatted result. For instance, suppose the Primary Grouping field is a boolean type, but the chart displays as Yes or No using the formatValue parameter. When using a link, the data in fv!selection for that field will be a boolean instead of a text. This applies when using some intervals that provide localized formatting too. For example, the DATE_TEXT interval returns the date as a text, like November 16, 2021. However, when you add a link for this field, the resulting value is the original date value 11/16/2021, not the localized text.
Once you store the currently selected chart data to a local variable, it's easy to use the selection for filtering, drilling, or other interactions on the page. See the recipe on how to configure a chart drilldown to a grid for an example on how to use the selection.
In addition to the groupings and measures, charts that use a record type as a source contain options for setting a batch size and sorting.
To limit the batch size, enter a value in the Data Limit field of the chart's Component Configuration. The data limit determines how many unique groupings are on the chart. The data limit is most useful when using a single grouping.
For example, if you want a chart to display the top 10 products by their sales, the limit of 10 will return the correct results. The data limit also determines the combination of groupings. So, if both a primary and secondary grouping are used, a data limit of 100 would return a total of 100 values (10 primary grouping values x 10 secondary grouping values). For this reason, we recommend using the maximum limit of 5,000 when using multiple groupings.
To sort the record data in the chart, use the Sort field of the chart's Component Configuration. Sort is used to determine which values display first on the chart.
When configuring the sort, always sort by a field used in the groupings or measures. When multiple groupings are used, it usually makes the most sense to sort by the primary grouping field.
To configure a chart's sort behavior:
Under Sort By, select the alias from the chart's grouping or measure.
In Design Mode, you can only sort by an alias. To add an alias, click the edit icon next to your measure and enter a value in the Alias field.
To configure the sort behavior in Expression Mode, use the a!sortInfo function.
Let's look at a few examples to understand how to configure the chart to display in a variety of scenarios.
In this example, say you have an Order record type with a many-to-one relationship with the Customer record type, and you want to see how many orders are associated with each customer. An easy way to display this data is using a column chart, and we'll use the Order record type as the source.
Since the Order record type has a relationship with the Customer record type, we can use the Name field from the Customer record type as the Primary Grouping value. Then we'll configure the Measure as count of ids to get the total number of orders for each customer.
In Design Mode, select the source of the record type and use the fields dropdown to select the record fields for grouping and measure. To select a related record field, hover over the relationship name listed in the fields dropdown.
In Expression Mode, we can provide the key fields in the data and config parameters (along with some other basic visual properties). To select a related record in an expression, use the recordType! domain and the relationships property to index into the fields in a relationship.
See Reference a Record Type in Expressions for more information.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
a!columnChartField(
data: recordType!Order,
config: a!columnChartConfig(
primaryGrouping: a!grouping(
field: recordType!Order.relationships.Customer.fields.name,
alias: "name_primaryGrouping"
),
measures: {
a!measure(
function: "COUNT",
field: recordType!Order.fields.id,
alias: "id_count_measure1"
)
},
dataLimit: 100
),
label: "Orders per Customer",
stacking: "NONE",
showLegend: true,
showTooltips: true,
labelPosition: "ABOVE",
colorScheme: "BERRY",
height: "MEDIUM",
xAxisStyle: "STANDARD",
yAxisStyle: "STANDARD"
)
Here's the final column chart:
Suppose you have a record type that contains a list of sales made for each product. A pie chart is a great way to view the contributions made by each product area.
Since the product area displays on each slice of the pie, this should be configured as the grouping field. The sales determine the size of each pie slice, so we should configure the measure as the sum of sales revenue.
In Design Mode, simply select the source record type and appropriate fields for grouping and measure.
In Expression Mode, we can provide the key fields in the data and config parameters (along with some other basic visual properties).
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
a!pieChartField(
label: "Sales by Product Area",
data: recordType!Sales,
config: a!pieChartConfig(
primaryGrouping: a!grouping(
field: recordType!Sales.fields.productArea,
alias: "productArea"
),
measures: a!measure(
function: "SUM",
field: recordType!Sales.fields.salesRevenue,
alias: "sum_salesRevenue"
),
dataLimit: 100,
sort: a!sortInfo(
field: "sum_salesRevenue",
ascending: false
)
),
labelPosition: "ABOVE",
height: "SHORT",
showDataLabels: true,
showAsPercentage: true,
seriesLabelStyle: "LEGEND",
colorScheme: "RAINFOREST",
style: "DONUT"
)
Here's the final pie chart:
In this example, we'd like to count the number of employees in the company grouped by their region and role. Since this can best be displayed as a stacked chart, we'll need to use multiple groupings. We should also only display active employees using a filter.
Since employees in each region have the same type of roles, we will use the region field as the Primary Grouping value. This is the value that will display on the axis of the chart. We'll use the role field as the Secondary Grouping, so the field labels display as stacked columns in the chart.
To configure this chart in Design Mode:
In Expression Mode, we can provide the key fields in the data and config parameters (along with some other basic visual properties). To display only the active employees, we'll use a!recordData() to create a filter.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
a!columnChartField(
data: a!recordData(
recordType: recordType!Employee,
filters: a!queryFilter(
field: recordType!Employee.fields.status,
operator: "=",
value: true
)
),
config: a!columnChartConfig(
primaryGrouping: a!grouping(
field: recordType!Employee.fields.region,
alias: "region"
),
secondaryGrouping: a!grouping(
field: recordType!Employee.fields.role,
alias: "role"
),
measures: a!measure(
function: "COUNT",
field: recordType!Employee.fields.id,
alias: "id_count"
),
dataLimit: 5000
),
label: "Column Chart",
stacking: "NORMAL",
showLegend: true,
showTooltips: true,
colorScheme: "MIDNIGHT"
)
Here's what the chart looks like:
Suppose you'd like to create a chart that displays the total number of cases created within each office over time. If you'd like to see long term trends, it makes sense to use an interval to group the data by month. Here's what the chart looks like:
To begin, in Design Mode:
Here's a sample expression for this chart:
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
a!lineChartField(
data: recordType!Case,
config: a!lineChartConfig(
primaryGrouping: a!grouping(
field: recordType!Case.fields.createdOn,
alias: "createdOn_month_primaryGrouping",
interval: "MONTH_TEXT"
),
secondaryGrouping: a!grouping(
field: recordType!Case.fields.office,
alias: "office"
),
measures: {
a!measure(
function: "COUNT",
field: recordType!Case.fields.id,
alias: "id_count"
)
},
dataLimit: 5000
),
label: "Cases by Month per Office",
showTooltips: true,
colorScheme: "RAINFOREST",
)
