Overview

The Write Records smart service is a powerful tool that allows you to easily insert or update data in your source system and then automatically sync those changes in Appian.

To write data using this smart service, you must use a record type that meets the following conditions:

• The record type uses a database as its source.
• The record type has data sync enabled.

You can only update records in one record type. To update a record and its related records, you must first update the base record, then update your related record with another Write Records node. Learn more about how to write records and related records.

Permissions needed

The user executing this node must have Viewer permissions to the selected record type in order to update or insert information. If the user does not, the node will fail and pause the process with an exception.

Properties

• Category: Data Services

• Icon:

• Assignment Options: Automated activity

Configuration options

This section contains configuration details specific to this smart service. For more information about common configurations, see the Process Node Properties page.

Setup tab

The Setup tab is where you'll configure the records to write, as well as any corresponding record events to write. If your record type does not have record events configured, you will only specify the records to write on this tab.

To configure the Setup tab:

1. In your process model, open a Write Records node.

2. Go to the Setup tab.

   Tip:  If you do not see the Setup tab in your node, replace the current node with a new Write Records node.

   

   

3. For Records Input, select a process variable or write an expression to specify the records to update. You can write a total of 50,000 records and events combined per node.
4. For Record Type, specify the record type you are updating. If you selected a process variable for your Records Input, this value is auto-populated by default with the selected record type.
5. If the selected record type has record events configured, specify when to write events for the selected record type under Choose when to write events:
   • Select Always to write events whenever the node executes.
   • Select Never to never write events whenever the node executes.
   • Select Only when… to specify the condition in which events are written. Then, click Edit Condition to write an expression defining the specific condition.

6. Select the values to write for the event. The field inputs are populated from the related Event History record type. You must map values for the following field inputs:

   Input	Value
   Event Type	Use the dropdown to select an event type or use an expression to specify the event type. The dropdown list is populated from the values in the Event Type Lookup record type. If you use an expression, you must use the event type identifier to indicate the type of event. Learn more about conditionally choosing event values.
   User	Use the dropdown to select the Process Initiator or a process variable of type User or Text, or use an expression to specify the user. If the event will be completed by an automated task, set the value to Null (Set as System Event).
   Timestamp	Use the dropdown to select Now (the date and time when the node is executed) or a process variable of type Date and Time, or use an expression to specify the date and time.

7. Click OK.
8. Click Save & Publish in the process model.

Data tab

The Data tab allows you to add additional inputs and store the resulting node outputs.

Node inputs

Node outputs

Usage considerations

Pass data using a record data type

To pass data into the Write Records smart service or function, you must use a record data type. This means that when you build a start form, you should use record types as your rule inputs or local variable, and any process variables should have a Type of Record Type.

If you're not passing data from a variable, you can use a record type constructor to pass in the record data type and the data to write in a script task.

Create new records

A new record will be created when one of the following is true:

• The source table is set to AUTO_INCREMENT the primary key column and you omit the primary key field or set it as null.
• The record type's primary key field is configured to use a sequence to generate key values and you omit the primary key field or set it as null.
• The source table is not configured to AUTO_INCREMENT or use a sequence and you provide a primary key value that does not exist.

Update existing records

To update existing data, the record must include a primary key value. For example, if you want to update an employee record, you must pass the employee's Id to the Write Records smart service.

When you update an existing record, only include the data you want to change. For example, imagine your Employee record type has fields for firstName and lastName, but you only need to update an employee's last name from Anderson to Kwan. You can pass just the record's identifier and the new lastName. Any fields not included in the update are ignored. In this example, the firstName field is not included, so the employee's first name remains unchanged.

1
2
3
4
recordType!Employee(
  recordType!Employee.fields.id: 233,
  recordType!Employee.fields.lastName: "Kwan"
)

Create and update related records

If you try to write related records in the Records Input, they will be ignored. For example, if you try to write orders and related order items, only orders will be written; the order items will be ignored.

Instead, to write related records, your process model must include a separate node for each record type and related record type you want to update. See an example below.

Unsupported fields in the Records Input

If you include custom record fields in the Records Input, they will be ignored. Custom record fields are defined by and exist only in Appian, so attempting to write values for them will have no effect.

Write records and events

If the record type selected in the Records Input has record events configured, you can write records and events at the same time, in the same node. This means that when the node executes, the record type will be updated, as well as the related Event History record type.

For example, let's say you configure a Write Records node to create a new order. If the Order record type has record events configured, you can specify that when a new order is created, a new event is also created to capture who created the order and when.

If you are updating multiple records in one node, then each new record will have a corresponding event created. For example, if you create order 1, order 2, and order 3, each order will have its own unique event created.

Conditionally choose event values

You can conditionally set the value for any event input, like Event Type, User, or Timestamp. This approach is useful when you use a single Write Records smart service to add and update data, or when you want to write different event values depending on a value in the record.

For example, when you generate Create and Update record actions at the same time, Appian will create a single process model to perform both actions. If you selected the Include Common Event Types option when you generated your event record types, the generated process model is automatically configured to write events correctly. As a result, "Created record" events are only written when a new record is created, and "Updated record" events are only written when a record is updated.

You may also want to write different event types depending on the value of a record. For example, let's say an order is updated to have a status of "Reviewed" so you want write an event called "Reviewed Order" instead of "Updated Order".

To conditionally set the event type for this example:

1. In your Write Records smart service, go to the Setup tab.
2. Next to Event Type, click Edit as Expression.

3. Write an expression using an if() statement and the order status value. Use the event type identifier to specify the event type. For example:

   1
   2
   3
   4
   5
   6
   7
   if(
    pv!record[recordType!Order.relationships.status.fields.name] = "Reviewed",
    /* The ID for the "Reviewed Order" event type */
    3,
    /* The ID for the "Updated Order" event type */
    2
   )
   

   Tip:  For easier maintenance, use a constant to specify the event type identifer instead of entering a static numerical value.

Write events without writing records

There are some scenarios where you may want to write an event without making any changes to the base record type.

For example, let's say you sent a feedback form to a customer after an order is shipped. If the customer doesn't respond to the form in five days, the process model will automatically send an email reminder.

To capture an event for the automatic email reminder, you can write directly to the Event History record type. To do this:

• Add a Script Task node and configure it with the data to pass to the Write Records node. Use a record type constructor to specify the event information to write to the Event History record type. For example:

  1
  2
  3
  4
  5
  6
   recordType!Customer Events(
       recordType!Customer Events.fields.recordId: pv!customer[recordType!Customer.fields.recordId],
       recordType!Customer Events.fields.timestamp: now(),
       recordType!Customer Events.fields.user: null, 
       recordType!Customer Events.fields.activityId: cons!CUSTOMER_EVENT_EMAIL_REMINDER_ID
   )
  

  Note:  A null user field indicates that this is an automated activity, it was not started by a specific user.

• Add a Write Records node and set the Records Input to a process variable or expression that evaluates to the Event History record type.

The updated process would look something like this:

Automatic refresh on record views

When you use this smart service on a record view, the view will automatically refresh with the latest record data.

This means that when your record view interface has a Record Action component that triggers the Write Records smart service, or a button component that triggers the a!writeRecords() function, the record view data will automatically refresh after the write occurs.

This refresh behavior only applies to data supplied by rv!record. If your record view definition uses rv!identifier or a local variable to supply the record data, the data will not automatically refresh; instead, you will need to configure additional refresh behavior.

a!writeRecords()

The Write Records smart service is available as an expression function that can be executed inside a saveInto on a Interface Component or as part of a Web API. When creating a web API, select the Write Records template to generate the object with this function automatically configured.

Syntax

a!writeRecords( records, onSuccess, onError )

Parameters

Possible errors

This section explains some of the common reasons the Write Records smart service might fail.

Configuration errors

Configuration errors can occur when the node setup or inputs are misconfigured and attempts to write to the source cannot succeed. Some examples of misconfiguration include:

• Using an unsupported record type (for example, a record type with a non-database source).
• An error in the related Event History record type or Event Type Lookup record type (for example, a field is deleted or the record type is deleted).
• An error in the expression calling the process model or the a!writeRecords() function.

Errors with the data source

Appian is unable to change data in the source in some circumstances. These situations include:

• Schema mismatches such as conflicting data types.
• Database connection problems like timeouts or invalid credentials.

If Appian is unable to write to the database, the record data will still remain available even if the smart service fails.

Errors syncing data changes

When Appian attempts to sync changes to the source data, the sync will fail when one of the following occurs:

• The record type's row limit is exceeded.
• A record field contains an invalid value.
• The record type's source filter is invalid (for example, if the source field is no longer available).

If Appian is unable to sync changes from the source, the record data will be unavailable. For more information on the different types of sync errors and steps to resolve these errors, see Troubleshooting Syncs.

Examples

The section walks you through the different ways you can use the Write Records smart service to create or update your record data.

Create a new record using a process model

Let's say you need to create a record action that allows users to add new customers in a retail application. The application contains a Customer record type, which stores all customer information. This record type also has record events configured, so you also want to capture an event whenever a new customer is created.

To create this action, you could:

1. Create an interface that allows users to enter the customer's information. Use the Customer record type as a rule input in the interface to save the entered data.
2. Create a new process model and configure the process model properties:
   • Go to the Variables tab and create a new process variable called customer. Set the Type to the Customer record type, and select the Parameter checkbox.
   

   

   • Go to the Process Start Form tab and select your interface. Map your customer process variable to the inputs from the interface.
3. Drag and drop a Write Records node into your process model.
4. Go to the Setup tab.
5. For Records Input, select the customer process variable. By default, the Record Type input populates with the Customer record type. Since the Customer record type has record events configured, the Write Events section can be set up.
6. For Choose when to write events for the Customer record type, leave the default Always. The smart service will always write an event whenever a new customer is created.
7. Select the values to write for the event:
   • For Event Type, select Created Customer.
   • For User, keep the default Process Initiator.
   • For Timestamp, keep the default Now.
   

   

8. Click OK.
9. (Optional) Configure a cancel flow.

The process model looks something like this:

When you start the process, Appian will only write data to the fields you specify. To give you a complete picture of your records, the node output will return all record fields and event field values—even those that were not updated.

Create a new record using an expression

Now let's say you want to create a form that updates your Status record type with a new status option for orders. Instead of building an entire process model to update this reference data, you can just build a form and use the a!writeRecords() function in the submit button. This will automatically write and sync any data provided in the form whenever a user clicks Submit.

To create the form in this example, you could:

1. Create a new interface using a form layout.
2. Create a local variable called status. This will store the new status value and pass it to the a!writeRecords() function.
3. Add components that allow users to add new values. In this example, you could add a Text field to allow users to add a new status label.
4. In the primary button of the form, add a saveInto parameter with the following values:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
    buttons: a!buttonLayout(
      primaryButtons: {
        a!buttonWidget(
          label: "Submit",
          saveInto: {
            a!writeRecords(
              records: {
                recordType!Status(
                  recordType!Status.fields.label: local!status
                )
              }
            )
          },
          submit: true,
          style: "PRIMARY"
          )
        },
      )

Now, whenever a user clicks Submit, a new record will be created in the Status record type.

Create records and related records

You cannot write records and related records in the same node. Instead, you will need to configure a separate Write Records node for each record type you want to update.

For example, let's say you need to create a new record action so users can enter information about new salespeople joining the company. To represent people, your application's data model includes three record types:

• Person, the base record type with data like firstName and lastName.
• Employee, a related record type for data like hireDate and dateOfBirth.
• Sales Person, a related record type with data for salesQuota and commissionRate.

This means you'll need to write to three different record types in order to add a new salesperson. As such, you will need three different Write Records smart services to update the different record types since you cannot use relationship references in the smart service.

The process model to create a new Sales Person will look something like this:

Let's break down what's happening:

• The process model uses an interface as a start form so new user data is passed to the process model and set as the initial values for three process variables (person, employee, and salesperson).

• The first Write Records node uses the person variable as the Records Input to create the new Person record.

  Since you need to reference the new person ID in order to create the related records in the Employee and Sales Person record types, you'll store the person ID in the output.

  

  

• The process variable person now includes the id generated for the person (in this example, the value is 20778).

  

  

• Next, a script task automatically sets the personId of the employee and salesperson variables to 20778 (the id of the Person record).

  

  

  

  

• The process writes the related Employee and Sales Person records in their own nodes, and they will be correctly associated with the Person.

  

  

Feature compatibility