fix(salesforce): translate metric expressions and validate field references - #404
Draft
saurabhdeshp wants to merge 6 commits into
Draft
saurabhdeshp wants to merge 6 commits into
saurabhdeshp wants to merge 6 commits into
Conversation
MetricMappingHandler.mapOssieToSalesforce() was a stub that discarded every metric mapping and returned without writing anything, so an Ossie model's metrics never appeared in the exported Salesforce semanticCalculatedMeasurements. Wire up the existing generic name/description mapping (mappings.yaml already declared metrics.name/metrics.description, unused until now), and add unwrapExpressions() to flatten each metric's expression into a Salesforce-compatible string. TABLEAU is preferred when present since that is what Salesforce/Tableau CRM speaks; a model authored without one falls back to its ANSI_SQL expression unresolved/untranslated, since actually resolving/rewriting into TABLEAU syntax depends on apache#222's still-open expression-language work, not this fix. A metric with neither dialect fails the conversion with an actionable error naming the metric, rather than being silently dropped. Fixes apache#399
This was referenced Sep 15, 2026
saurabhdeshp
marked this pull request as ready for review
September 15, 2026 23:22
saurabhdeshp
marked this pull request as draft
September 16, 2026 01:55
Use JSqlParser with a shared typed expression pipeline for derived fields and dependent metrics. Preserve semantic identities through environment bindings, validate relationships and final model references, and reject unsupported expressions or ambiguous metadata before export. Validation: 442 tests passed with no skips, Apache RAT passed, and packaged CLI export, import, and physical-binding smoke checks passed.
Defer physical bindings, derived fields, metric dependencies, extension changes and whole-model validation. Keep JSqlParser-backed metric parsing, typed Tua generation and targeted field/relationship reference checks. Consolidate the compiler around metric-specific components and retain focused semantic and integration regressions. Validation: 254 tests passed with no skips, Apache RAT passed, and the packaged CLI export smoke check passed.
… tests Normalize bounded unary chains and redundant grouping before JSqlParser, while preserving operand validation, quotes, function arguments and limits. Reject malformed null predicates and tuple arguments explicitly. Fix ELSEIF evaluation and signed-zero equality, distinct counting and zero-divisor checks in the independent semantic test helper. Add focused compatibility, semantic and rejection regressions. Validation: 280 tests passed with no skips; Apache RAT passed. Packaged CLI checks cover successful conversion and failure without partial output.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Implements the initial metric-expression scope of #403 on top of #402. Supported OSI
SNOWFLAKEandANSI_SQLformulas compile to Tableau NextTua; invalid types, aggregation or references fail with the metric name. PreferredTABLEAUformulas are validated too.Supports aggregates, arithmetic, searched
CASE, Boolean comparisons, null handling andABS/ROUND/CEIL/FLOOR. For example,SUM(orders.profit) / NULLIF(SUM(orders.revenue), 0)becomes a bracketed Tua ratio with anIFnull guard.The PR is limited to metric conversion: 16 changed files, 1,220 production additions. Deployment bindings, derived-field compilation, metric dependencies, date/string functions, extension changes and whole-model validation are deferred. The existing field/relationship mapping pipeline stays in place.
Review the implementation in this order:
MetricMappingHandlercalls the compiler with one resolver per model.SqlMetricExpressionParseruses JSqlParser 5.3;TuaMetricExpressionParserhandles the bounded native syntax. Both return the smallMetricExpressionAST.MetricExpressionTranslatorchecks types/aggregation before emitting Tua.MetricFieldResolververifies declared and exported identities, physical bindings, and the enabled relationship paths needed by metrics. It rejects derived expressions misclassified as physical columns and changed join-key pairs; it does not rewrite the relationship mapper or prove join grain.JSqlParser is used under its Apache-2.0 option, with its unused JMH dependency excluded. No Python runtime is required.
Validation: JDK 21
mvn -DrequireSalesforceSchema=true clean verifypassed 280 tests with zero failures/errors/skips; Apache RAT and packaged-JAR export smoke checks passed. Tests include repeated unary operators, redundant grouping, malformed modifier/tuple rejection, ordered ELSEIF evaluation, signed-zero handling, composed formulas, nulls/empty inputs/duplicates/zero denominators, quoting, parser bounds, types, relationship correspondence, and CLI failures without partial output.Draft: native Tableau Next validation remains pending. Local evaluation and schema tests do not prove authoring API acceptance or native results. #403 stays open for native validation and later milestones.
Refs #399 and #403. Merge #402 first; its commit and authorship are preserved. Extension-preservation work belongs there. The CLI error-output line overlaps #286.