Dataflow Conversion Mapping

During conversion, dataflow functionality is mapped to similar recipe functionality. Consider the mapping specifications when evaluating the conversion of your dataflows to recipes.

Dataflow Transformations

At a glance dataflow to recipe conversion mapping.

Dataflow	Conversion	Recipe
append		Append
augment		Join
computeExpression		Transform (Custom formula)
computeRelative	see parameters	Transform (Custom formula with multiple rows)
delta
digest		Input
dim2mea		Transform (Dimension to measure)
edgemart		Input
export	see parameters	Output (as .csv)
filter		Filter
flatten		Transform (Flatten)
prediction		Discovery Predict
sfdcDigest	see parameters	Input and Transform
sfdcRegister		Output
sliceDataset		Transform (Drop columns)
update		Update

append
Dataflow Parameter	Conversion	Supported in Recipe
sources		Append
enableDisjointedSchemaMerge		Append (Allow schema merge)

augment
Dataflow Parameter	Conversion	Supported in Recipe
operation		Join (lookup)
left		Join (lookup)
left_key		Join (lookup)
right		Join (lookup)
relationship		Join (lookup)
right_select		Join (lookup)
right_key		Join (lookup)

computeExpression
Dataflow Parameter	Conversion	Supported in Recipe
mergeWithSource		Transform (Drop columns)
saqlExpression		Transform (SQL in custom formula)
format		Transform (Custom formula)
fiscalMonthOffset		Use Data Prep Date Configurations See Data Prep Fiscal Configuration
isYearEndFiscalYear		Use Data Prep Date Configurations See Data Prep Fiscal Configuration
firstDayOfWeek		Use Data Prep Date Configurations See Data Prep Fiscal Configuration
precision		Transform (Custom formula)
scale		Transform (Custom formula)
defaultValue		Transform (Custom formula) The defaultValue in dataflows support function calls. For example, `now()`. In recipes, function calls aren’t supported and the defaultValue must be a literal.

computeRelative
Dataflow Parameter	Conversion	Supported in Recipe
partitionBy		Transform (Custom formula with multiple rows)
orderBy		Transform (Custom formula with multiple rows)
computedFields.name		Transform (Custom formula with multiple rows)
computedFields.label		Transform (Custom formula with multiple rows)
computedFields.description
computedFields.expression.sourceField		Transform (Custom formula with multiple rows)
computedFields.expression.offset		Transform (Custom formula with multiple rows)
computedFields.expression.default		Transform (Custom formula with multiple rows)
computedFields.expression.saqlExpression		Transform (Custom formula with SQL) Self-referencing formulas don’t convert and aren’t supported. For example, using computeRelative to create a field “MyNewField”, which has SAQL expression referencing previous(MyNewField). Achieve similar results using recipes. For example, if using computeRelative in dataflow to: Aggregate based on data grouping, try using recipe Aggregate. Concatenate a column’s multiple text values into a column based on a data grouping, try using recipe Join lookup with multivalue and Transformation custom formula with `array_join`. Generate row numbers, try using recipe Transformation custom formula with `row_number`. Be used in a filter to keep only the most recent records based on a data grouping - such as the last activity record for an account, try using recipe Aggregate and Filter.
computedFields.expression.type		Transform (Custom formula with multiple rows)
computedFields.expression.scale		Transform (Custom formula with multiple rows)

Delta transform isn’t supported in Recipes. To calculate changes over time, consider using a transformation node with windowing functions in multiple row custom formulas.

digest
Dataflow Parameter	Conversion	Supported in Recipe
connectionName		Input
object		Input
fields		Input

dim2mea
Dataflow Parameter	Conversion	Supported in Recipe
measureDefault		Transform (Dimension to measure)
measureType

To extract all columns in a dataset, use the Include All Fields option in the Input node.

edgemart
Dataflow Parameter	Conversion	Supported in Recipe
alias		Input

export
Dataflow Parameter	Conversion	Supported in Recipe
source		Output (to .csv)
target
userId

filter
Dataflow Parameter	Conversion	Supported in Recipe
filter		Filter
saqlFilter		Filter (SQL filter expression) SAQL `like` operator may not be converted (depending on complexity) and isn’t supported. Use recipe Transformation custom formula with `startsWith`, `endsWith`, or `contains`.

flatten
Dataflow Parameter	Conversion	Supported in Recipe
include_self_id		Transform (Flatten)
self_field		Transform (Flatten)
parent_field		Transform (Flatten)
multi_field		Transform (Flatten)
path_field		Transform (Flatten)

prediction
Dataflow Parameter	Conversion	Supported in Recipe
predictionDefinitionName		Discovery Predict
predictionColumnName		Discovery Predict
predictionColumnLabel		Discovery Predict

sfdcDigest
Dataflow Parameter	Conversion	Supported in Recipe
object		Input
incremental
fullRefreshToken
fields		Input
filterConditions		Filter (Filter logic)
complexFilterConditions		Use Filter (SQL expression)

sfdcDigest
Dataflow Field Attributes	Conversion	Supported in Recipe
name		Input
defaultValue		Use Transform (Custom formula with `case` or `coalesce`)
type		Use Transform (Replace or Custom formula)
fiscalMonthOffset		Use Data Prep Date Configurations See Data Prep Fiscal Configuration
isYearEndFiscalYear		Use Data Prep Date Configurations See Data Prep Fiscal Configuration
firstDayOfWeek		Data Prep Date Configurations See Data Prep Fiscal Configuration
isMultiValue		Use Transform (Custom formula multivalue functions) See Multivalue Functions for Formulas
multiValueSeperator		Use Transform (Custom formula multivalue functions) See Multivalue Functions for Formulas
precision		Use Transform (Edit attributes) See Edit Attributes Transformation
scale		Use Transform (Edit attributes) See Edit Attributes Transformation

sfdcRegister
Dataflow Parameter	Conversion	Supported in Recipe
alias		Output
name		Output
source		Output
rowLevelSharingSource		Output
rowLevelSecurityFilter		Output

sliceDataset
Dataflow Parameter	Conversion	Supported in Recipe
source		Transform (Drop columns)
mode		Transform (Drop columns)
fields		Transform (Drop columns)

update
Dataflow Parameter	Conversion	Supported in Recipe
left		Update
right		Update
left_key		Update
right_key		Update
update_columns		Update

Dataflow SAQL

Dataflows use SAQL and recipes use SQL. During conversion, SAQL is changed to SQL whenever possible.

Warning SAQL date filtering, relative date ranges, and references to date epochs are not supported. Recommendation is to evaluate supported solutions to replace dataflow SAQL date capabilities.

In recipes, try using simple operators with Dates or DateTimes such as +,-,=,<,>,<=,>=. For example, add hours by using (DateTime) + (TimeInterval).

Below are instances when SAQL is not automatically converted to SQL, and the recommended SQL alternative. In recipes, use SQL in a Transform custom formula.

Dataflow (SAQL)	Recipe (SQL)
DateField_sec_epoch	to_unix_timestamp(DateField)
DateField_day_epoch	to_unix_timestamp(DateField) / (60 * 60 * 24)
DateField_Year	format_number(year(DateField), '0000')
DateField_Quarter	format_number(quarter(DateField), ‘0')
DateField_Month	format_number(month(DateField), '00')
DateField_Week	format_number(weekofyear(DateField), '00')
DateField_Day	format_number(day(DateField), '00')
DateField_Hour	format_number(hour(DateField), '00')
DateField_Minute	format_number(minute(DateField), '00')
DateField_Second	format_number(second(DateField), '00')
toDate()	to_timestamp()
date_to_epoch()	to_unix_timestamp()
month_first_day(toDate('CloseDate_sec_epoch'))	date_trunc('month', CloseDate)
month_last_day(toDate('CloseDate_sec_epoch'))	date_trunc('month', CloseDate) + INTERVAL 1 month - INTERVAL 1 day
week_first_day(toDate('CloseDate_sec_epoch')) This function always returns Sunday of the week	date_trunc('week', CloseDate) This function always returns Monday of the week
Other first_day or last_day functions	Use date_trunc on the different parts: 'year', 'quarter', 'month', 'week'.
toString(toDate('CloseDate_sec_epoch'), "yyyy-MM-dd")	date_format(CloseDate, 'yyyy-MM-dd')
date_to_string()	date_format(CloseDate, 'yyyy-MM-dd')
date_diff("year", toDate(StartDate_sec_epoch), toDate(EndDate_sec_epoch))	year(EndDate) - year(StartDate)
date_diff("month", toDate(StartDate_sec_epoch), toDate(EndDate_sec_epoch))	months_between(EndDate, StartDate)
date_diff("quarter", toDate(StartDate_sec_epoch), toDate(EndDate_sec_epoch))	(year(EndDate)-year(StartDate))*4 + quarter(EndDate) - quarter(StartDate)
date_diff("week", toDate(StartDate_sec_epoch), toDate(EndDate_sec_epoch))	floor(datediff(EndDate, StartDate) / 7)
date_diff("day", toDate(StartDate_sec_epoch), toDate(EndDate_sec_epoch))	datediff(EndDate, StartDate)
date_diff("hour", toDate(StartDate_sec_epoch), toDate(EndDate_sec_epoch))	(to_unix_timestamp(EndDate) - to_unix_timestamp(StartDate))/(60*60)
date_diff("minute", toDate(StartDate_sec_epoch), toDate(EndDate_sec_epoch))	(to_unix_timestamp(EndDate) - to_unix_timestamp(StartDate))/(60)
date_diff("second", toDate(StartDate_sec_epoch), toDate(EndDate_sec_epoch))	to_unix_timestamp(EndDate) - to_unix_timestamp(StartDate)