Moving from Lextegrity to Quantum data extracts
For customers migrating an existing Lextegrity data extract integration to Quantum.
Table of Contents
hThis guide is for administrators, data engineers, and reporting teams moving an existing Lextegrity reporting integration to Quantum. It covers changes to extract files and downstream reports, and the steps for switching production reporting to Quantum.
The application migration and reporting integration are related workstreams. The extract limitations described here concern what is delivered in reporting files; they do not define the scope of data migrated into the Quantum application.
File names, columns, identifiers, status values, and some reporting detail change. Update and test your integration before switching it over. For file formats, dataset keys, and loading examples, see Quantum data extracts: Format and delivery guide. New Quantum customers should start there.
What your integration needs to change
| Area | Lextegrity | Quantum migration action |
|---|---|---|
| Dataset catalog | 22 extract files | Map to 21 datasets, including merged files, new child datasets, and one file without a replacement. |
| Delta selection | Rows loaded in a rolling 24-hour window | Apply explicit insert, update, and delete changes based on comparison with the last successful delivery. Do not substitute timestamp filtering. |
| Deletions | Generally not emitted; some identity files used history rows | Handle delete rows by the complete dataset key. Distinguish hard deletes from soft-removal timestamps. |
| Full extracts | Requested manually | Use the automatic full on the first of each month alongside that day’s delta. The full resets the delivery baseline. |
| Delivery structure | One uncompressed CSV per file | Read one gzip-compressed CSV extract file per record set (dataset), with JSON manifests, in monthly YYYY-MM folders. Each record set is delivered as a single file, not multiple parts. A delivery can contain multiple record sets. The filename identifies full or delta. |
| Schedule | Daily task at 00:30 UTC | Use the regional 01:00 local cutoff: America/New_York or Europe/Paris. Data and filename timestamps use UTC. See the format guide for delivery targets. |
| Connection | Existing Lextegrity SFTP access | Open Data extracts in the admin panel and use SFTP access to view the Quantum connection details and manage SSH connection setup. |
| Monitoring | Existing file and retry assumptions | Validate each dataset file against its manifest and monitor all expected datasets. Datasets can retry independently. |
Map your Lextegrity files
This table identifies affected pipelines. Use the complete column catalog and sample deliveries supplied by Case IQ for parser implementation.
| Lextegrity dataset | Quantum dataset | Migration consideration |
|---|---|---|
| requests | records | Status and decision fields change; forms_of_support and sub_type move to response rows. |
| activities | record_units | Primary responses and benefit/spend lines; entered and normalized amounts are separate. |
| request_activity_parties | record_unit_parties | One participant share, with requested, approved, and actual amounts. |
| responses | record_responses | Typed values; one row per selected option. File answers move to attachments. |
| actions | record_actions | Multiple event sources; key by action_id + action_source. |
| request_activity_approval_steps + request_recipient_approval_steps | approval_steps | Two files merge; stage_scope and participant_party_id distinguish scope. |
| request_file_attachments | attachments | File-link metadata from answers, records, comments, and tasks, distinguished by link_source. |
| tasks | tasks + task_parties | One row per task; linked parties move to a child dataset. |
| parties | parties | One person or organization; only the active FMV version. |
| party_addresses | party_addresses | Adds address_id; state becomes region_code and region_text. |
| party_details | party_attributes | Typed customer-defined attributes; employer is also represented as a relationship. |
| party_relationships | party_relationships | Each directed relationship is stored once, rather than in both directions. |
| users | users | One derived account status and one manager relationship replace legacy fields. |
| user_roles | user_roles | Effective-dated assignments; removals typically arrive as updates that set is_active to false. Apply the update by assignment_id and exclude inactive assignments from current-access reporting. |
| role_privileges | role_capabilities | Capability keys replace privilege codes; revise the mapping. |
| user_groups | user_groups | Membership derives from rules; the admitting rule is part of the key. |
| user_privileges | user_capability_exceptions | Effective-dated per-user capability grants and denies. |
| risk_scoring | risk_scores | Risk bucket, brief, and rationale replace numeric scores. |
| risk_scoring_criteria | risk_score_factors | Factors cited by the model, not all legacy criteria. |
| risk_scoring_adjustment | No replacement | Manual risk-score adjustments are not provided. |
| audit_logs | extract_log | Delivery counts, size, checksum, schema version, and retry linkage. |
Rebuild identifiers and joins
Quantum uses new UUID identifiers. Migrated rows retain legacy references where defined: legacy_record_id, legacy_party_id, legacy_user_id, and legacy_task_reference. Build an ID mapping table (sometimes called a crosswalk) to connect old identifiers to new ones. Use Quantum IDs for ongoing processing. A legacy task reference is not a Quantum task ID.
Example: map a request and update its joins
The following fictional example uses shortened IDs for readability; actual identifiers are UUIDs.
| Source | Field | Example value |
|---|---|---|
| Lextegrity requests | request_id | LEX-123 |
| Quantum records | legacy_record_id | LEX-123 |
| Quantum records | record_id | Q-789 |
| Quantum record_units | record_id | Q-789 |
- Match the old
requests.request_idtorecords.legacy_record_idto reconcile the migrated disclosure. - Store the mapping from LEX-123 to Q-789 in your reporting system.
- Join Quantum units to the record using
record_units.record_id = records.record_id. Both contain Q-789. - Apply future record updates and deletes using Q-789. Do not join Quantum child rows to the old Lextegrity ID or the human-readable reference.
Submitters, approvers, and event actors generally use *_user_id. Join these to users.user_id; use documented party references for joins to parties.party_id. User and party IDs are not interchangeable.
Understand what one row represents
- Tasks: one row per task, rather than a repeated task for each response. Linked parties are in task_parties.
- Relationships: one directed relationship, rather than reciprocal duplicate rows.
- Multi-select answers: one row per selected option; apply changes using answer_id.
- Approver groups: a user admitted by two rules has two rows. The key is group_id + user_id + rule_id.
- Workflow events: the key is action_id + action_source.
- Approvals: the key is stage_id + approver_user_id. Preserve a stage with no approver; its approver identifier is empty.
Different row counts do not necessarily mean missing data. Reconcile business records and amounts, and aggregate child data before joining it to other child datasets. See the worked disclosure example in the format and delivery guide for avoiding duplicated totals.
Update important field mappings
| Lextegrity field or assumption | Quantum replacement | What to review |
|---|---|---|
| customer_id | tenant_slug | Update tenant filters and labels. |
| request_id, formatted_request_id | record_id, reference | Keep the new UUID, human-readable reference, and legacy ID distinct. |
| completion_datetime | decided_at, closed_at, or completed_at | Select by lifecycle meaning; revalidate completion and cycle-time metrics. |
| activities.amount and currency_code | value_amount/value_currency; normalized_amount/normalized_currency | Distinguish entered from reporting-currency amounts. Record total_amount sums non-removed spend lines. |
| responses.response_data | value_text, value_number, value_date, value_code, value_ref and companion fields | Use the question type; date ranges include value_date_end, references have names, and file answers move to attachments. |
| Currency answers | value_number + value_code; converted_amount + converted_currency_code | The first pair records the entered amount/currency; the second records its comparison-currency value. These differ from normalized spend-line fields. |
| question_code, question_default_text, response_type | question_key, question_label, question_type | Retain form-version context when mapping answers. |
| actioned_by_party_id, action_datetime, comment | actor_user_id, occurred_at, comment_body | Change actor joins and interpret action_source. |
| workflow_step_id, request_step_order, request_step_name | stage_id, stage_sequence, stage_name | Two approval files merge; stage decided_at and member approved_at describe different events. |
| Aggregate modified_datetime filter | Dataset timestamps and explicit change_type | Consume delivered deltas rather than reconstructing a 24-hour window. |
For example, a fictional currency answer entered as 100 EUR and converted to 110 USD has value_number = 100.0000, value_code = EUR, converted_amount = 110.0000, and converted_currency_code = USD. This illustrates field usage, not a current exchange rate. Do not add monetary answers to spend totals merely because they contain amounts.
Replace status and risk logic
Record status
Replace Lextegrity status filters with Quantum’s statuses:
open, info_requested, on_hold, approved, rejected,
closed, canceled, expired, update_pending, retiredMigrated records retain the Lextegrity code in legacy_status for reconciliation. Use Quantum status for ongoing reporting. Validate each report’s mapping rather than assuming a universal one-to-one translation.
Use records.status for record approval. A primary response unit may remain open after the record is decided. Tasks, users, parties, and stages have their own status vocabularies. User account status is active, locked, deactivated, or deleted.
Risk reporting
risk_scores provides low, med, high, or critical buckets, with a brief and rationale. Numeric scores, maximum scores, and manual adjustments are not provided. risk_score_factors contains model-cited factors, not the complete legacy criteria model. Only risk-enabled record types have score rows.
Replace numeric thresholds and adjustment-based metrics. Have report owners approve replacement metrics and explain why historic numeric scores and Quantum buckets are not directly comparable.
Account for reporting differences
- Migrated action history in extracts: derived status transitions are included, not the complete Lextegrity per-action history. Agree how historical reporting will remain available.
- Voided actions: superseded approximates voided only for information requests replaced by later requests. They are not interchangeable. requires_mitigation has no equivalent.
- Managers: the four approver-manager slots and HR-manager field are not exported. Quantum has one manager relationship.
- Approval references: request_step_id, workflow_user_id, and workflow_collaborator_id have no direct equivalents. Review dependent joins.
- Tasks: response-request and completion details are on the task, replacing the Lextegrity task-response model. Legacy details, template_code, automated_task_template_id, and product_code have no direct equivalents; origin identifies automated tasks.
- FMV: extracts include the active fair-market-value version, not one row per historical version.
- Other removed fields: group codes, relationship descriptions, user-agent/IP audit columns, and internal attachment storage paths are not provided.
- Audit scope: extract_log reports deliveries, not the internal application audit trail or the identity of the Case IQ operator requesting a run.
- Anonymity: submitter and author identities are blank where required; event actors are masked when they are the anonymous submitter. Preserve these rules downstream.
- Approver reassignment: process delivered seat deletions; do not assume every former approver remains as an inactive seat.
Plan and validate the switch to Quantum
The production switch is often called cutover. Complete these steps before pointing production reports at Quantum data.
- Inventory reports and integrations. List every consumed file, column, key, join, status filter, risk calculation, and historical report. Assign an owner.
- Prepare access and reference material. Open Data extracts in the admin panel and use SFTP access for connection details and SSH setup. Obtain the Quantum schemas and sample full/delta/manifest files, and agree the schedule and switch date.
- Build and test. Update compressed CSV and manifest handling, map files and fields, build ID mapping tables, and revise status/risk logic.
- Load a full baseline. Record its cutoff. Reconcile representative records, decisions, answers, participant amounts, tasks, parties, and relationships. Explain expected count differences.
- Test changes. Exercise inserts, updates, hard deletes, soft removals, header-only results, and composite keys. Verify that replay does not duplicate data.
- Test recovery. Detect missing manifests or dataset files and failed checksums. Recover a failed ingestion even when delivery succeeded; do not assume a later delta repeats missed changes.
- Test the monthly refresh. Follow the full/delta loading example. A full already contains the state at its cutoff; do not add the same-cutoff delta on top of that refreshed current-state table. Keep separate checkpoints for any change-history store.
- Obtain reporting sign-off. Have report owners approve reconciliations, revised metrics, and historical limitations. Agree recovery arrangements before the switch.
- Switch and monitor. Validate all expected datasets and logs, reconcile initial production loads, and retain the agreed historical reporting access.
Common migration questions
Can we reuse our existing pipeline?
Some connection or ingestion components may be reusable, but review parsing, schemas, keys, joins, status filters, risk calculations, and delete handling. Do not assume the pipeline will work unchanged.
Why do row counts change?
Tasks, relationships, multi-select answers, and rule-derived memberships represent rows differently. Compare business meaning and amounts rather than expecting identical file counts.
Where are the ongoing file and delivery instructions?
Use Quantum data extracts: Format and delivery guide for naming, CSV types, keys, manifest validation, and full/delta loading examples.
What should we include in a support request?
Provide the dataset, cutoff, full/delta kind, manifest or filename, processing error, and failed count, size, or checksum checks. For reconciliation issues, include the legacy and Quantum references where available, using only the data needed to investigate.
Keep inbound HR delivery separate
Reporting extracts move data out of Quantum. HR sync moves employee data into Quantum. The Data management → SFTP HR intake screen configures inbound files; use the extract-specific connection for outbound reporting. See Configure HR data sync and SFTP intake for the inbound workflow.