[{"id":"cassandraadvanced-001","courseId":"cassandra-advanced","domain":"data-modeling","topic":"Per-cell overhead in size estimates","prompt":"Your team sizes a partition with the on-disk formula. For each cell it adds a metadata term, and it uses 8 bytes per cell as that value. Which statement about the per-cell figure is correct for Cassandra 3.0 and later?","hint":"Look at the word the page uses for the 8-byte figure, and what the storage format changes per cell.","explain":"Each stored value costs bytes beyond its payload, such as timestamps. A sizing model that charges one average for that overhead depends on how uniform it really is across values.","source":"https://cassandra.apache.org/doc/latest/cassandra/developing/data-modeling/data-modeling_refining.html","interaction":{"type":"single"},"options":[{"id":"a","text":"Every cell carries exactly 8 bytes of metadata, so the estimate is exact for every table.","whyNot":"The page calls 8 bytes a typical estimate and says the size formula is only an approximation."},{"id":"b","text":"The per-cell metadata varies with the data type and with whether custom timestamps or TTLs are set; 8 bytes is typical.","why":"The page says metadata varies by data type and by custom timestamp or TTL settings, and offers 8 bytes as a typical estimate."},{"id":"c","text":"Metadata is counted once per row, not per cell, so it does not grow with the number of values.","whyNot":"The metadata term is multiplied by the number of values, so it is charged for every cell."},{"id":"d","text":"Only cells with a custom TTL store metadata; ordinary cells store none at all.","whyNot":"Cells carry metadata such as timestamps even without custom values; the page says the amount varies, not that it vanishes."}],"answer":"b","explanation":"In Cassandra 3.0+ storage, per-cell metadata varies with the data type and with custom timestamp or TTL settings; 8 bytes is a typical estimate, so the size formula is an approximation.","concept":"In Cassandra 3.0+ storage, per-cell metadata varies with the data type and with custom timestamp or TTL settings; 8 bytes is a typical estimate, so the size formula is an approximation.","rev":1,"updatedAt":"2026-10-26","kind":"scenario","takeaways":["Size formulas are approximations; per-cell overhead varies, so model it as a range, not a constant."]},{"id":"cassandraadvanced-002","courseId":"cassandra-advanced","domain":"data-modeling","topic":"Single-replica size estimates","prompt":"Judge this statement: the on-disk size estimate for a partition already counts every replica the cluster keeps, so total table capacity needs no replication multiplier.","hint":"Think about how many copies the estimate describes, and what the replication setting decides.","explain":"A size formula describes one thing, while capacity planning covers every place data lives. Consider what the replication setting controls and whether the formula's output reflects it.","source":"https://cassandra.apache.org/doc/latest/cassandra/developing/data-modeling/data-modeling_refining.html","interaction":{"type":"true-false"},"options":[{"id":"a","text":"True","whyNot":"It sounds plausible, but the page says the estimate counts only a single replica, so it covers one copy."},{"id":"b","text":"False","why":"The page says the estimate counts a single replica; total capacity must multiply it by the number of partitions and the replicas set by the replication strategy."}],"answer":"b","explanation":"The per-partition size formula counts a single replica. Total capacity multiplies that size by the number of partitions and by the replicas specified by the keyspace's replication strategy.","concept":"The per-partition size formula counts a single replica. Total capacity multiplies that size by the number of partitions and by the replicas specified by the keyspace's replication strategy.","rev":1,"updatedAt":"2026-10-26","kind":"scenario","takeaways":["Multiply single-replica estimates by partitions and replicas before planning cluster capacity."]},{"id":"cassandraadvanced-003","courseId":"cassandra-advanced","domain":"data-modeling","topic":"Frozen user-defined types","prompt":"A type is created in the hotel keyspace, as in the code. A table in the reservation keyspace needs an address column of that type. Which approach works in CQL?","hint":"Ask where a type name is valid, and what the page says about who can use a type.","explain":"A user-defined type bundles named fields under one name. Think about the scope in which a type name is valid, and whether a schema element can be shared across separate namespaces.","source":"https://cassandra.apache.org/doc/latest/cassandra/developing/cql/types.html","interaction":{"type":"single"},"options":[{"id":"a","text":"Write hotel.address as the column type, since the keyspace prefix lets the table reach it.","whyNot":"A keyspace prefix only sets where a type is created; the type still works only in its own keyspace."},{"id":"b","text":"Add the frozen keyword to the column, which lets the type be used across keyspaces.","whyNot":"The page says a type can only be used in the keyspace where it was created, and no keyword changes that rule."},{"id":"c","text":"Declare a new address type inside the reservation keyspace, then use it in that keyspace's tables.","why":"A type is bound to the keyspace that creates it and can only be used there, so reservation needs its own."},{"id":"d","text":"Nothing is needed, because user-defined types are shared across the whole cluster.","whyNot":"The page says a type is intrinsically bound to its keyspace, so it is not cluster-wide."}],"answer":"c","explanation":"A user-defined type is bound to the keyspace where it is created and can only be used there; a keyspace prefix only controls where it is created.","concept":"A user-defined type is bound to the keyspace where it is created and can only be used there; a keyspace prefix only controls where it is created.","rev":1,"updatedAt":"2026-10-26","kind":"scenario","code":"CREATE TYPE hotel.address (\n  street text,\n  city text\n);","takeaways":["A UDT lives in one keyspace; declare it again in every other keyspace that needs it."]},{"id":"cassandraadvanced-004","courseId":"cassandra-advanced","domain":"data-modeling","topic":"Tuples as anonymous types","prompt":"A table has a column declared as tuple<int, text>. Which TWO statements about how this column can be written and updated are correct?","hint":"Think about how tuples differ from UDTs, which have named fields.","explain":"Composed types differ in how their parts can be changed. Compare a type with named fields that can be edited one at a time with a type whose parts stand together.","source":"https://cassandra.apache.org/doc/latest/cassandra/developing/cql/types.html","interaction":{"type":"multiple","tellCount":true},"options":[{"id":"a","text":"The column is always frozen, so the frozen keyword is not written in the definition.","why":"The page says a tuple is always frozen without needing the keyword, unlike collections and UDTs."},{"id":"b","text":"A partial UPDATE can change only the text element, the same way it can change one UDT field.","whyNot":"The page says it is not possible to update only some elements of a tuple; the whole tuple is written."},{"id":"c","text":"Changing only the text element is impossible, so the whole tuple must be written again.","why":"The page says a tuple cannot be updated element by element, so the full value is written."},{"id":"d","text":"A literal may leave out trailing elements, and they are set to NULL.","whyNot":"The page says a tuple literal must supply every declared element; NULLs must be written explicitly."},{"id":"e","text":"Each element has a name, like a UDT field, so queries can address an element by name.","whyNot":"The page calls tuples anonymous types with anonymous fields, so their elements have no names."},{"id":"f","text":"Every element must share one type, as the items of a list do.","whyNot":"The page says tuple elements can be of different types."}],"answer":"ac","explanation":"A tuple is always frozen and has no field names. Its elements cannot be updated one at a time, and a literal must give a value for each declared element.","concept":"A tuple is always frozen and has no field names. Its elements cannot be updated one at a time, and a literal must give a value for each declared element.","rev":1,"updatedAt":"2026-10-26","kind":"scenario","takeaways":["Tuples are written whole: plan updates as full-value rewrites, not field edits."]},{"id":"cassandraadvanced-005","courseId":"cassandra-advanced","domain":"data-modeling","topic":"Physical design build order","prompt":"A team turns a finished logical hotel model into a physical one. Put these steps in the order the physical data modeling page describes them.","hint":"Ask which activity needs concrete types already chosen before it can produce any useful result.","explain":"Physical design is a cycle of choices and checks. Some checks only yield meaningful results once the choices they test are fixed.","source":"https://cassandra.apache.org/doc/latest/cassandra/developing/data-modeling/data-modeling_physical.html","interaction":{"type":"sort","items":[{"text":"Define the logical data model that the physical model starts from.","why":"The page says creating the physical model is a process that begins once a logical data model is defined."},{"text":"Walk through each logical model table, assigning a type to each item.","why":"The page says you walk through the logical model tables assigning types to each item, and only after that analyze the model."},{"text":"Analyze the model with size calculations and by testing how it works.","why":"The page says that after assigning data types, you analyze the model by performing size calculations and testing how it works."},{"text":"Adjust the design based on the findings from that analysis.","why":"The page says you may make some adjustments based on your findings, which come from the analysis step."}]},"options":[],"answer":"o:abcd","explanation":"The physical modeling page walks the logical tables to assign types, then analyzes the model with size calculations and testing, and then may adjust the design based on those findings.","concept":"The physical modeling page walks the logical tables to assign types, then analyzes the model with size calculations and testing, and then may adjust the design based on those findings.","rev":1,"updatedAt":"2026-10-26","kind":"scenario","takeaways":["Assign types first, then measure the design with size calculations and tests before accepting it."]}]