Denodo Platform 9 Certified Developer Professional flashcards

35 cards. Read the term and the definition here. The deck at the top of this page is optional.

Base view vs data source

A data source holds connection metadata only. A base view wraps a table or operation from that source and exposes it as a queryable VDP view with its own schema, metadata, and delegation settings.

Delegation capability

A per-source, per-database declaration of which functions and operators the underlying system can evaluate. Disabling a capability forces Denodo to compute it locally, affecting both cost and result semantics.

Credential pass-through

A configuration where the end user's own credentials are used to connect to the source instead of a shared stored account. Row-level security at the source then applies per user.

Connection pool sizing

Each data source has initial, minimum, and maximum pool sizes. The maximum must stay below the source database's connection limit, or connections fail intermittently when the pool saturates.

Hash join

A join strategy that builds an in-memory hash table on the smaller input and probes it with the larger one. Chosen by the optimizer when one side is small enough to fit in memory.

Cross-source join

A join whose inputs come from different data sources. It cannot be delegated and always executes in the Denodo server, making it the main consumer of server memory and CPU.

Execution trace

A per-query report showing the generated SQL sent to each source, rows and time per node, and whether each node was pushed down or executed locally. The primary tuning instrument.

Pushdown blocker: implicit cast

When union branches or join keys have mismatched types, Denodo inserts casts. Those casts often cannot be delegated, turning a pushed-down operation into a local one. Align types explicitly.

Union type alignment

All branches of a union must expose compatible column types. Declaring matching types in each branch avoids implicit casts and keeps the union delegated to the source.

Association

A documented logical relationship between fields of two views, with a cardinality such as one-to-many. It does not change execution; it records the model and enables join suggestions.

Tag

Name/value metadata attached to views, columns, databases, or sources. Used for classification such as PII or sensitivity level, and available to policies and catalog search.

Controlled tag vocabulary

A server-level tag definition with a fixed set of allowed values. It keeps classification consistent and lets security policies match on tag values instead of free text.

Column-level privilege

A grant that permits a role to see specific columns of a view. Queries referencing forbidden columns fail or exclude them, depending on how the privilege is configured.

Row restriction

A filter injected into every query on a view, typically using the current user's identity, so the same view returns different rows per user without duplicating the view.

Data masking

A mask type applied to a column (full, partial, hash, random, custom expression). The masked value is computed in Denodo, so clients never receive the original.

Unmask privilege

The privilege that lets a role see original values instead of masked ones. Administrators usually hold it, so mask testing must be done with a non-privileged user.

Additive privileges

Denodo authorization is role-based and grants are additive. There is no explicit deny that overrides a grant, so restriction must come from withholding grants or from row restrictions.

REST view publication

Exposing a view as an HTTP service with a chosen method, input parameters, output format, and paging. Paging must be set explicitly or large results return in a single response.

OData query options

Published OData entity sets support $filter, $select, $orderby, $top, $skip, and $expand. Unsupported operations return an error rather than being silently ignored.

GraphQL field selection

Clients request exactly the fields they need, and nested selections follow associations. Field-level authorization still applies, so a forbidden field is rejected even when requested.

Service authentication scope

Published services run as the authenticated user, so row restrictions and masks apply to API calls exactly as they do to direct SQL clients.

Cache modes

Off, partial (cache only fully delegated results), or full. A cached entry is reused only when the same query and parameters recur before the entry expires.

Cache invalidation

Entries expire by TTL, can be invalidated manually, or can be invalidated when the source changes. Without a TTL, stale results persist until explicit invalidation.

MPP acceleration

Pre-computing view data into an external parallel engine so repeated analytical queries avoid the source. Configured per view with a refresh schedule.

Statistics and join order

The optimizer's join order and strategy depend on row-count estimates. Missing or stale statistics on a large join is a common cause of memory errors and poor plans.

Scheduler job and task

A job contains one or more tasks; a task runs a query or a data movement step. A cron-like schedule or a dependency on another job triggers execution.

Incremental load watermark

A monotonically increasing column such as last_modified plus a stored high-water mark. Each run selects rows above the mark and updates the mark only after a successful load.

Load modes

Append inserts new rows, overwrite truncates and reloads, and incremental inserts new rows and optionally updates changed rows using a key and a watermark column.

Scheduler repository

The database where Scheduler stores jobs, schedules, and execution history. History records rows read and written, duration, and errors, forming the audit trail for data movement.

Solution Manager revision

A versioned export of a set of model elements from a source environment. Revisions are compared against a target to produce a reviewed deployment plan of creates, updates, and deletes.

VQL export

The text format for model elements, usable for scripted or partial deployment. It does not manage revisions or rollback the way Solution Manager does.

Deployment property substitution

Environment-specific values such as hostnames and credentials are overridden at deployment time. Hard-coding development values is a common cause of production deployment failure.

VDP metadata repository

The store holding all VDP model elements and security configuration. An external RDBMS is recommended for production so backup, HA, and growth are manageable.

Distributed installation

Installing the VDP server, web container, Scheduler, and Solution Manager on separate hosts instead of together. Required for scale and for isolating the query engine from web traffic.

Backup scope

A complete backup includes the metadata repository, configuration files, and custom extensions such as drivers. View definitions are recoverable only from these or from a saved VQL export.