Recipe and dataflow transformation choices
Which recipe transformation or node meets a requirement (bucket, split, dates, formula across rows, types and attributes, joins, append, aggregate, built-in machine learning), the dataflow transformation that matches it, and why the recipe preview is only a sample.
Given business/user requirements, perform data transformations in dataflows/recipes.
~8 min read · 17 practice questions ·
What the exam expects
- Pick the recipe transformation for a value-shaping need: Bucket, Split, Format Dates, Extract, Date and Time, type conversion, Edit Attributes, Drop Columns.
- Build calculated columns with Formula, within one row or across rows with partition and order.
- Combine inputs with the right join type, Append or Aggregate, and work round the join key-pair limit.
- Choose the built-in machine learning transformation (Detect Sentiment, Cluster, Predict Missing Values, Time Series Forecasting, Discovery Predict).
- Map a requirement to the dataflow transformation (computeExpression, computeRelative, sliceDataset, augment, Delta) when a dataflow must stay.
Key terms: Transform node (the recipe node that holds a sequence of transformations); multiple-row formula (a Formula with window functions, Partition By and Order By); lookup (a join that keeps one output row per left row); Append (stacks rows of inputs that share columns); preview sample (the subset of rows the recipe editor shows while you build).
Key facts
- Detail
- Aggregate, Cluster, Detect Sentiment, Predict Missing Values, Split, Time Series Forecasting and more
- Detail
- Delta
- Detail
- computeExpression (same row), computeRelative (across rows), sliceDataset (drop), augment (lookup)
- Detail
- Two values per Split at a delimiter
- Detail
- Multiple row formula with Partition By and Order By
- Detail
- Labels, API names, precision, scale, text length, date formats
- Detail
- Up to 5 key pairs
- Detail
- Lookup keeps one row per left row; joins keep all matches
- Detail
- Up to two dimension columns as columns
- Detail
- A sample of the input, sized differently on the two pages; the run processes every row
Sources for this section (8)
- Why Use Recipeshelp.salesforce.com (opens in a new tab)
- Transformations for Data Prephelp.salesforce.com (opens in a new tab)
- Formula Transformationhelp.salesforce.com (opens in a new tab)
- Prepare Your Datatrailhead.salesforce.com (opens in a new tab)
- Join Operationshelp.salesforce.com (opens in a new tab)
- Aggregate Nodehelp.salesforce.com (opens in a new tab)
- Preview Resultshelp.salesforce.com (opens in a new tab)
- Set Up the Previewhelp.salesforce.com (opens in a new tab)
- Why can a left join change a dataset's grain when a lookup on the same key does not?
- When does a cross join make sense, given that it takes no keys?
- What length does a text calculated column get if you set none?
- Why might a node after a join show no sample rows while the full run returns data?
The questions are free to read. Answers with their sources and self-grading are included with the pack.
Keep reading with the CRM Analytics and Einstein Discovery Consultant pack
- Core concepts
- Exam traps
- Worked scenario
The pack includes:
- All 26 lessons of the CRM Analytics and Einstein Discovery Consultant guide, with reading progress
- Interactive decision walks and step-by-step sequence steppers
- Practice exams purchase adds inline bank questions and topic practice (17 questions for this lesson)
- An official source cited for every claim
- An AI tutor that answers from the guide
- Choose guide-only, practice-only, or save $4 with both