Skip to main content
Hooks are a powerful feature that allow you to programmatically validate, transform, and enhance data during the import process. They integrate seamlessly with field validation and results handling to give you complete control over your data pipeline.
Looking for simpler validation options? Check out field validators for built-in validation rules, or see the validation guide for best practices.
The diagram below shows the general flow for Dromo. Column hooks run first and only once after the “column matching” stage. After these run, row hooks will run for each record. Row hooks will also run on a record after it is updated by the user. If you only want to run a row hook on init or on change, you can configure it accordingly.

​
Column hooks​

Column hooks run on a specified column at the beginning of the data review step. They are run before the row hooks and will only run once during the import process. Column hooks are ideal for batch operations like API lookups or data enrichment for entire columns.
To register column hooks, use registerColumnHook() in the JavaScript SDK or the columnHooks prop in the React component.
Column hooks are called with a single argument, values, with an array of objects. Each object corresponds to one value in the column, and has two parameters.
​
index
number
The row index of the column value, starting with 0
​
value
string
The raw value as imported
Column hooks must return an array of objects in a similar format. Column hooks may return the value async, wrapped in a promise.
​
index
number
required
The row index of the column value. This should be unchanged from the input.
​
value
string
The new value for this cell. If omitted, the value is left unchanged.
​
info
InfoMessage[]
An array of messages you would like to add to this cell.For more details, see Info Messages.If omitted, no messages are added.
Given the following imported data:
NameEmail
Michelle Doe[email protected]
Jane Doe[email protected]
Steve Jones[email protected]
Wayne Gretzky[email protected]
A column hook registered on the Email field would be passed the following data:
[
  { value: "[email protected]", index: 0 },
  { value: "[email protected]", index: 1 },
  { value: "[email protected]", index: 2 },
  { value: "[email protected]", index: 3 },
];
This column hook could return something like this:
[
  {
    index: 0,
    value: "[email protected]",
    info: [
      {
        message: "Automatically corrected email domain",
        level: "info", // should be "info", "warning" or "error"
      },
    ],
  },
  // we can omit cells that don't need any changes or messages
  {
    index: 2,
    // Don't need to supply value if we're not changing it
    info: [
      {
        message: "Email not found",
        level: "error",
      },
    ],
  },
];

const fields = [
  {
    label: "Email",
    key: "email",
  },
];

const settings = {
  importIdentifier: "Column Hooks Server Validation",
};

const user = {
  id: "12345",
};

const dromo = new DromoUploader("FRONTEND_API_KEY", fields, settings, user);

dromo.registerColumnHook("email", function (values) {
  return values.map((row) => {
    const newRow = {
      index: row.index,
      value: "0" + row.value,
      info: [
        {
          message: "Prepended 0 to value",
          level: "info",
        },
      ],
    };
    return newRow;
  });
});

​
Row hooks​

Row hooks run validation on each record (row) of data and return the record with new data and/or user messages. Row hooks come in two variants: standard and bulk.
To register row hooks, use registerRowHook() or registerBulkRowHook() in the JavaScript SDK, or the rowHooks/bulkRowHooks props in the React component.
Row hooks are run once in "init" mode after the column matching step and before the user begins the review step, and are run with the full dataset. Row hooks are run again in "update" mode after each time the user changes any data, with only the changed rows.

​
Standard row hooks​

Standard row hooks are invoked once for every row on init, and once for every row that changed on update. Standard row hooks are called with two arguments.
​
record
object
​
index
number
The index of this row
​
row
object
The row object will have keys matching the field key of each mapped field. Each of these keys will contain an object with data and metadata of that field in the given row.
​
value
string
The value of this field at this row, as displayed to the user
​
resultValue
string | boolean | number | null
The value of this field at this row, as it will appear in the result data
​
info
InfoMessage[]
An array of any info messages you have previously added to this cell
​
selectOptions
{ label: string, value: string }[]
Overridden selectOptions previously added to this cell
​
mode
"init" | "update"
Mode will be "init" if running before the start of the review step, and "update" if running due to a user change.
​
manyToOne
{ value: string, resultValue: string | boolean | number | null, info: InfoMessage[], selectOptions: { label: string, value: string }[] }[]
If the field is manyToOne, every mapped cell for this row appears as an array.
Row hooks should return the updated record in a similar format. Row hooks may return the value async, wrapped in a promise.
​
row
object
required
The row object should contain keys matching the field key of each mapped field. If a mapped field is omitted, it will remain unchanged.Each field key object may have the following entries:
​
value
string
The new value for this row
​
info
InfoMessage[]
New messages for this row. Messages returned will replace any previously set messages. If info is omitted, messages will remain unchanged.
​
selectOptions
{ label: string, value: string }[]
Select option overrides for this cell. If provided, the cell will use these select options for transforming values and validating instead of those set when the field was declared.
When fetching external data, we recommend that you make any API calls in a column hook and then validate that the data remains correct using a row hook in update mode.
const fields = [
  {
    label: "Email",
    key: "email",
  },
  {
    label: "State",
    key: "state",
    selectOptions: [{ label: "Alabama", value: "AL" }, { label: "Alaska", value: "AK" }],
  },
];

const settings = {
  importIdentifier: "Row Hooks Example",
};

const user = {
  id: "12345",
};

const dromo = new DromoUploader("FRONTEND_API_KEY", fields, settings, user);

dromo.registerRowHook(function (record, mode) {
  const newRecord = record;
  if (record.index < 10 && mode === "init") {
    newRecord.row.email.value = "0" + newRecord.row.email.value;
    newRecord.row.email.info = [
      {
        message: "Prepend 0 to value",
        level: "info",
      },
    ];
    newRecord.row.state.selectOptions = [
      { label: "Alberta", value: "AB"}, 
      { label: "British Colombia", value: "BC"}
    ];
  }
  return newRecord;
});

​
Bulk row hooks​

Bulk row hooks are invoked a single time on init with the full dataset, and a single time on update with all of the changed rows. Bulk row hooks are a good fit if you need to perform potentially long-running operations that can be optimized by processing all changes in a single function call. Bulk row hooks work exactly like standard row hooks, except the first parameter is an array of record objects instead of a single record object. Similarly, bulk row hooks must return an array of record objects, preserving the index parameter of each record. 
const fields = [
  {
    label: "Email",
    key: "email",
  },
];

const settings = {
  importIdentifier: "Row Hooks Example",
};

const user = {
  id: "12345",
};

const dromo = new DromoUploader("FRONTEND_API_KEY", fields, settings, user);

dromo.registerBulkRowHook(function (records, mode) {
  return records.map((record) => {
    const newRecord = record;
    if (record.index < 10 && mode === "init") {
      newRecord.row.email.value = "0" + newRecord.row.email.value;
      newRecord.row.email.info = [
        {
          message: "Prepend 0 to value",
          level: "info",
        },
      ];
    }
    return newRecord;
  });
});

​
Row delete hooks​

Row delete hooks are called when a row (or rows) is removed from the table by the user. Row delete hooks are called once for each row that was removed.
​
record
object
​
index
number
The index of this row
​
row
object
The row object will have keys matching the field key of each mapped field. Each of these keys will contain an object with data and metadata of that field in the given row.
​
value
string
The value of this field at this row, as displayed to the user
​
resultValue
string | boolean | number | null
The value of this field at this row, as it will appear in the result data
​
info
InfoMessage[]
An array of any info messages you have previously added to this cell
dromo.registerRowDeleteHook(function (record) {
  console.log("Deleted row index: " + record.index);
});

​
Step hooks​

Step hooks are callbacks that are invoked at different parts of the import process. In addition to fetching metadata about the import at that time, you can use them to call mutation functions like addField.
To register step hooks, use registerStepHook() in the JavaScript SDK or the stepHooks prop in the React component. See the using addField guide for advanced use cases.

​
UPLOAD_STEP hook​

This hook is called when the user has completed uploading a file and provides you with a 20 row data preview of that data. The function is called with two arguments.
​
instance
DromoUploader
The Dromo Uploader instance which you can call mutation methods on
​
previewData
string[][]
An array of arrays containing the first 20 rows of raw uploaded data
dromo.registerStepHook("UPLOAD_STEP", function (instance, data) {
  instance.addField({
    label: "Full Name",
    key: "fullName",
  });
  console.log(data.filename, data.dataPreview);
});

​
REVIEW_STEP hook​

This hook is triggered just before the final review screen loads (and before any of the row and column hooks). The function is called with two arguments.
​
instance
DromoUploader
The Dromo Uploader instance which you can call mutation methods on
​
reviewStepData
object
​
rawHeaders
string[] | null
The raw headers from the uploaded file. If no file was uploaded (e.g. manual entry was used), this will be null.
​
headerMapping
{ [header: string]: string }
An object whose keys are the file headers and values are the field keys of the field the header was mapped to
​
fields
object
An object whose keys are the field keys and values are objects containing metadata about that field
​
fileHeader
string | null
The header this field was mapped to. null if no file was uploaded.
​
fileHeaderIndex
number | null
The column header index this field was mapped to. null if no file was uploaded.
​
isCustom
boolean
Whether this field was added by the user as a custom field
​
manyToOne
boolean
Indicates if this field is manyToOne.
​
fileHeaders
string[]
The headers that were mapped to the field if it is manyToOne.
​
fileHeaderIndexes
number[]
The indexes of columns that were mapped to a manyToOne field.
dromo.registerStepHook("REVIEW_STEP", function (instance, data) {
  instance.addField({
    label: "Full Name",
    key: "fullName",
  });
  console.log(data.rawHeaders, data.headerMapping, data.fields);
});

​
REVIEW_STEP_POST_HOOKS hook​

This hook is triggered during the review step after all of the initial row and column hooks have run (and after any `REVIEW_STEP` hooks have run). It can be handy for doing batched updates to the data based on aggregated results from row or column hooks. The function is called with two arguments.
​
instance
DromoUploader
The Dromo Uploader instance which you can call mutation methods on
​
reviewStepData
object
​
headerMapping
{ [header: string]: string }
An object whose keys are the file headers and values are the field keys of the field the header was mapped to
​
fields
object
An object whose keys are the field keys and values are objects containing metadata about that field
​
fileHeader
string | null
The header this field was mapped to. null if no file was uploaded.
​
fileHeaderIndex
number | null
The column header index this field was mapped to. null if no file was uploaded.
​
isCustom
boolean
Whether this field was added by the user as a custom field
dromo.registerStepHook("REVIEW_STEP_POST_HOOKS", function (instance, data) {
  instance.addInfoMessages([
    {
      rowIndex: 5,
      fieldKey: "month",
      level: "info",
      message: "Sales exceed previous month",
    },
  ]);
  console.log(data.headerMapping, data.fields);
});

​
REVIEW_STEP_PRE_SUBMIT hook​

This hook is triggered during the review step right after the user clicks ‘Finish’ but before beforeFinish runs. It can be useful for customizing the ‘Are you ready to submit?’ dialog with info summarizing the changes and impact of the import data. The function is called with two arguments.
​
instance
DromoUploader
The Dromo Uploader instance which you can call mutation methods on
​
reviewStepData
object
​
headerMapping
{ [header: string]: string }
An object whose keys are the file headers and values are the field keys of the field the header was mapped to
​
fields
object
An object whose keys are the field keys and values are objects containing metadata about that field
​
fileHeader
string | null
The header this field was mapped to. null if no file was uploaded.
​
fileHeaderIndex
number | null
The column header index this field was mapped to. null if no file was uploaded.
​
isCustom
boolean
Whether this field was added by the user as a custom field
dromo.registerStepHook("REVIEW_STEP_PRE_SUBMIT", function (instance, data) {
  instance.setConfirmationMessage(
    "<div>Adding 5 new contacts. Modifying 20 existing contacts.</div>",
    {
      submitButtonText: "Accept changes",
      cancelButtonText: "Cancel",
    }
  );
  console.log(data.headerMapping, data.fields);
});

​
Working with info messages​

Dromo offers the ability to add info messages to data cells to enable custom validation and provide other feedback to the user. Info messages are a key part of the validation system and work alongside field validators. You can add and update info messages via column hooks, row hooks, and the updateInfoMessages method in step hooks.
For simple validation scenarios, consider using built-in field validators before implementing custom validation with info messages.
Info messages are provided as objects with these parameters.
​
message
string
required
The message to be displayed to the user in the review screen when they hover over the corresponding cell
​
level
"info" | "warning" | "error"
default:"error"
The nature of the message.
  • A cell with an "error" message is highlighted in red. It is considered a validation error and impacts the user’s submission options based on invalidDataBehavior
  • A cell with a "warning" message is highlighted in yellow
  • A cell with an "info" message is highlighted in blue

​
The Dromo Uploader instance​

Step hooks provide access to the Dromo Uploader instance, which exposes methods to alter the import.

​
instance.addField​

Adds a field to the schema after a step hook has completed. By default, the field will be added to the end of the fields. You may optionally provide a position object specifying where in the list of fields to insert the new one.
​
field
Field
required
A field spec for the new field to add
​
position
{ before: string } | { after: string }
Specifies where in the list of fields to add the new one.If provided, must be an object of the form { before: fieldKey } or { after: fieldKey } where fieldKey is the key of a mapped field. If no mapped field with the given key exists, the new field will be added to the end.
For an example of concatenating or splitting fields, see the Guide.

​
instance.removeField​

Removes a field after a step hook has completed. The field will no longer be visible on the data review screen nor included in the result data. Any validation errors or info messages will be discarded. Can be used in conjuction with addField to concatenate columns and remove the originals, i.e. combine firstName + lastName into fullName and remove firstName and lastName from the result data.
​
fieldKey
string
required
Key of the field to be deleted.
For an example of concatenating or splitting fields, see the Guide.

​
instance.updateInfoMessages​

Updates the messages for targeted cells, overwriting any previous messages.
​
updates
object[]
required
An array of objects with an object for each cell to update the messages on
​
rowIndex
number
required
The 0-indexed position of the row you are targeting
​
fieldKey
string
required
The key of the field you are targeting
​
manyToOneIndex
number
If manyToOne is enabled on a field, you can use this property to select which mapped column shall receive the updates.
​
messages
InfoMessage[]
required
An array of infoMessages which will be set for the cell at the specified row and field.An empty array will clear existing messages from the cell.
In this case, if you had previously set a message on the email field of row 3 or the phone field of row 8, these messages would be replaced with the new ones.
instance.updateInfoMessages([
  {
    rowIndex: 3,
    fieldKey: "email",
    messages: [{ level: "warning", message: "Email has not been confirmed" }],
  },
  {
    rowIndex: 8,
    fieldKey: "phone",
    messages: [{ level: "error", message: "Phone number not in database" }],
  },
]);

​
instance.addRows​

Adds rows to the dataset. Returns an array of row IDs for the rows that were just added.
​
rows
object[]
required
Rows to be added to the dataset.
​
index
number
Index where this row will be inserted. If omitted, row will be added to end of the dataset.
​
row
object
The row object should contain keys matching the field key of each mapped field. If a mapped field is omitted, it will remain unchanged.Each field key object may have the following entries:
​
value
string
The new value for this row
​
info
InfoMessage[]
Info messages for this row.
​
selectOptions
{ label: string, value: string }[]
Select option overrides for this cell. If provided, the cell will use these select options for transforming values and validating instead of those set when the field was declared.

​
instance.removeRows​

Removes rows rows the dataset.
​
rowIds
string[]
required
Rows to be removed. Row IDs can be read on any row or column hooks.

​
instance.setConfirmationMessage​

Allows customization of the messaged displayed to the user after they click ‘Finish’. Content can be plain text or HTML.
​
messageHTML
string
required
Message for finish confirmation alert modal. Text or HTML for embedding links or media.
​
options
{ submitButtonText?: string, cancelButtonText?: string }
Optionally override text for submit and cancel buttons.