QSDA flashcards
QSDA · Build recall with a focused review session.
32 cards. Read the term and the definition here. The deck at the top of this page is optional.
- Synthetic key
- An implicit key table Qlik creates when two or more fields form a composite association between two tables; it multiplies rows and should be replaced by a hashed or concatenated key.
- Circular reference
- Occurs when three or more tables are linked in a loop. It is resolved with a link table, by renaming one key field, or by removing the redundant association.
- Star schema
- A model with one central fact table connected to flat dimension tables by single key fields, so every dimension is exactly one association away from the facts.
- Snowflake schema
- A star schema whose dimension tables are normalized into further sub-dimension tables, increasing join depth and association count and usually reducing performance versus a star.
- Qualify prefix
- A script prefix that adds the table name to every field name in the following LOAD, preventing unintended associations between tables that share generic names such as ID or Date.
- Concatenate prefix
- Appends the rows of a new table onto an existing table with matching field names; NOConcatenate prevents automatic concatenation and FORCE overrides mismatched field names.
- Preceding load
- Stacked LOAD statements where the upper LOAD reads the result of the LOAD below it; the script executes bottom-up, letting one source read feed several transformations.
- Resident load
- Loads data from a table already in memory instead of the original source, avoiding repeated database or file access and enabling stepwise aggregation and re-use.
- ApplyMap function
- Maps values using a table created with Mapping LOAD; unmatched values return the original input unless a third default argument is supplied in the call.
- Buffer prefix
- Caches the result of a slow LOAD, typically a database query, to a local file so subsequent reloads read the cache instead of re-querying the source system.
- Binary load
- Loads the complete data model of another app without running its script; useful for chaining apps, but the imported tables can no longer be modified.
- Incremental load pattern
- Load only new or changed records based on a stored maximum key, concatenate them with the existing QVD, then store the combined table back to that QVD.
- alt() function
- Returns the first non-NULL value among its arguments, used to coalesce or default fields during cleansing without writing deeply nested if() statements.
- NullAsValue statement
- Script statement that converts NULL values in the listed fields into a chosen value on load, so they become a visible, selectable dimension value.
- Exists() function
- Returns true when a field value already exists in a loaded table; most often used inside a WHERE clause to filter records during incremental extraction.
- Granularity
- The level of detail of a table as defined by its key fields; joining tables of different grain duplicates rows and inflates every sum and count.
- IntervalMatch prefix
- Links discrete numeric values to intervals held in another table; it produces a synthetic composite key that must be dropped after the match.
- Cartesian product
- The full row combination produced when two tables have no common field or an ambiguous association; record counts multiply and aggregations are silently corrupted.
- AutoNumber function
- Replaces distinct text key values with sequential integers, shrinking the symbol table and speeding up associations and aggregation in the in-memory model.
- Drop Field statement
- Removes fields from the in-memory model at a chosen point in the script, cutting RAM use and association overhead before the table is stored.
- Set analysis
- Expression syntax applying a selection-independent filter set at chart evaluation time; evaluated once per chart, so it outperforms row-level if() logic.
- QVD file
- A Qlik-native columnar compressed file holding one table plus an XML header with load and field metadata; loading it is far faster than re-querying the source.
- Optimized QVD load
- A QVD load keeps maximum speed only when written as LOAD * FROM file.qvd with no renamed fields, no calculated fields and no WHERE clause.
- SECTION ACCESS statement
- Script block defining user, group, password and access rights; it reduces data per user at reload time and must include at least one ACCESS=ADMIN entry.
- ADMIN versus USER access
- ADMIN rows see all data and are required for the app to open; USER rows are subject to the reduction defined by the field values listed for them.
- OMIT directive
- A section access directive that hides a named field from specified users, so they can neither select it nor view it on any sheet.
- Section access wildcard
- An asterisk placed in a section access field value grants that user every value of the field, effectively bypassing reduction for that column.
- Initial data reduction setting
- A Qlik Sense app setting that must be enabled for section access to take effect; when off, the section access block is ignored and all users see all data.
- QVD layer architecture
- A staged pipeline of raw extract QVDs, cleansed conformed QVDs and app-specific mart QVDs, with published apps loading only from the mart layer.
- Task chaining
- A QMC scheduler dependency that starts a downstream reload when the upstream task completes, automating the extract, transform and app refresh sequence.
- Stream
- A Qlik Sense container that organizes published apps and governs access through security rules; users only see the streams they have rights to.
- App promotion
- Moving an app between Dev, Test and Prod by exporting the .qvf from the QMC and importing it, or by using the Qlik command line tooling.