It's All Relative(ly Relational)
Edgar Codd proposed this data model in 1970, organizing data into tables of rows.
What is the relational model?
SQL, Cypher, and SPARQL are this type of query language—you state what you want, not how to get it.
What is declarative?
Storing human-meaningful info in just one place and referencing it everywhere by ID is this process.
What is normalization?
A graph's two object types are edges and these—also known as nodes or entities.
What are vertices?
R, Pandas, and Spark all support this table-like structure that data scientists "wrangle."
What is a DataFrame?
A résumé's positions and education form this branching structure that JSON makes explicit.
What is a tree?
Storing a whole document as one continuous string gives this read-performance advantage when you need the whole thing.
What is locality?
Many people live in one region—making the region_id field an example of this relationship type.
What is a many-to-one relationship?
This restrictive, OLTP-focused query language lets a client request a JSON response shaped exactly like its query.
What is GraphQL?
Turning a movie's genre into 1s and 0s spread across one column per possible value is this encoding technique.
What is one-hot encoding?
Splitting a document-like structure into multiple relational tables earns this cumbersome-sounding name.
What is shredding?
MongoDB's answer to SQL's GROUP BY uses this multi-stage construct with operators like $match and $group.
What is the aggregation pipeline?
Denormalization can be viewed as a form of this, since you must set up a process to keep redundant copies updated.
What is derived data?
The Turtle syntax encodes data in this Semantic Web data model whose name abbreviates "Resource Description Framework."
What is RDF?
Treating an immutable, append-only log of events as the source of truth is this pattern.
What is event sourcing?
The document model struggles to reference these directly—forcing you to say "the second item in the list of positions."
What are nested items?
This term describes interpreting data's structure only when reading it—the database analog of dynamic type checking.
What is schema-on-read?
At the center of a star schema sits this table, where each row represents an event like a customer's purchase.
What is the fact table?
Because events are never modified, the log can also serve as this, prized in regulated industries.
What is an audit log?
Codd's original relational model allowed these JSON-like nested values, decades before SQL formally added JSON.
What are nonsimple domains?
This clumsy SQL construct expands a 4-line Cypher query into 31 lines.
What is a recursive common table expression (WITH RECURSIVE)?
Folding all dimension data into denormalized columns of the fact table yields this approach, abbreviated ???.
What is one big table?
Researchers running DNA sequence-similarity searches rely on this specialized genome database.
What is GenBank?
Encrypting personal data with a deletable key—so an immutable log can still satisfy a GDPR deletion request—is this technique.
What is crypto-shredding?