myep/docs/buspronet-schema/schema-refresh-queries.md

Schema refresh & Zustieg-pricing verification queries

Superseded by docs/buspro-database-reference.md §12, which carries the refresh queries and the verification log. Retained only for the 2026-08-04 run detail; safe to delete once that is no longer of interest.

Read-only queries to (a) check whether the June schema dump still matches production after a BusPro update, (b) regenerate it if not, and (c) resolve the open Abfahrtsortzuschlag (Rückfahrt) question from the database side.

Current dump: BusPronet_20260601-1517.csv (columns), BusPronet_20260601-1518.csv (foreign keys), exported 2026-06-01.

All queries are SELECT-only. Run in order — Query 1 decides whether 2–4 are needed at all.

---

1. Drift check — run this first

Cheapest, highest signal. If it returns nothing, the June dump is still accurate and the re-export can be skipped.

SELECT s.name AS schema_name, o.name AS object_name, o.type_desc,
       o.create_date, o.modify_date
FROM sys.objects o
JOIN sys.schemas s ON s.schema_id = o.schema_id
WHERE o.is_ms_shipped = 0
  AND (o.create_date > '2026-06-01' OR o.modify_date > '2026-06-01')
ORDER BY o.modify_date DESC;

Note: modify_date moves on metadata changes, so a long list is expected after a version upgrade. What matters is whether anything under Zustieg*, Leistung*, Buchung* appears.

Result 2026-08-04. Every object reports create_date 2026-06-03 — the database was restored/rebuilt that day, so create_date is a restore artifact and the > '2026-06-01' filter matches everything. Use modify_date after the restore instead. Only two genuine clusters exist:

• 2026-06-11 — six new tables, all absent from …-1517.csv: Reichweite, ReichweiteZeitraum, ReichweiteZeitraumVonBis, ReichweiteRestFarbe, LadungTankungBetriebshof, ConnectViafintechSlip.
• 2026-07-02 — modified: Tachostand, GutscheinAktion, AdressePersonalZusatz, BlockPlanung, Block, LeistungZusatz, ConnectViafintechSlip, TourPlanung.

Nothing under Zustieg*, ZustiegPreis, ArtZustiegPreis, LeistungBefoerderung or BuchungTNZuschlag changed after the restore. The June dump is accurate for the booking and boarding areas; the only staleness is the six fleet/payment tables above. For future runs, filter on o.modify_date > '2026-06-03' to skip the restore baseline.

---

2. Columns — regenerates …-1517.csv

Column order and names match the existing header exactly, so the output is a drop-in replacement. Export semicolon-separated with quoted values, same settings as the June file.

SELECT s.name        AS schema_name,
       t.name        AS table_name,
       c.name        AS column_name,
       ty.name       AS data_type,
       c.max_length,
       c.precision,
       c.scale,
       c.is_nullable,
       c.is_identity,
       dc.definition AS default_definition
FROM sys.tables t
JOIN sys.schemas s          ON s.schema_id = t.schema_id
JOIN sys.columns c          ON c.object_id = t.object_id
JOIN sys.types ty           ON ty.user_type_id = c.user_type_id
LEFT JOIN sys.default_constraints dc ON dc.object_id = c.default_object_id
ORDER BY t.name, c.column_id;

3. Foreign keys — regenerates …-1518.csv

SELECT fk.name AS foreign_key_name,
       cs.name AS child_schema,  ct.name AS child_table,  cc.name AS child_column,
       ps.name AS parent_schema, pt.name AS parent_table, pc.name AS parent_column
FROM sys.foreign_keys fk
JOIN sys.foreign_key_columns fkc ON fkc.constraint_object_id = fk.object_id
JOIN sys.tables  ct ON ct.object_id = fkc.parent_object_id
JOIN sys.schemas cs ON cs.schema_id = ct.schema_id
JOIN sys.columns cc ON cc.object_id = fkc.parent_object_id
                   AND cc.column_id = fkc.parent_column_id
JOIN sys.tables  pt ON pt.object_id = fkc.referenced_object_id
JOIN sys.schemas ps ON ps.schema_id = pt.schema_id
JOIN sys.columns pc ON pc.object_id = fkc.referenced_object_id
                   AND pc.column_id = fkc.referenced_column_id
ORDER BY ct.name, fk.name, fkc.constraint_column_id;

4. Primary keys & indexes — the documented gap

bpn-connect-query-reference.md records that PK/index data was expected but missing, and that identity flags are not a reliable PK substitute here. This closes it. Suggested filename: BusPronet_<date>-pk-index.csv.

SELECT s.name  AS schema_name,
       t.name  AS table_name,
       i.name  AS index_name,
       i.type_desc,
       i.is_primary_key,
       i.is_unique,
       i.is_unique_constraint,
       ic.key_ordinal,
       c.name  AS column_name,
       ic.is_included_column,
       ic.is_descending_key
FROM sys.indexes i
JOIN sys.tables  t ON t.object_id = i.object_id
JOIN sys.schemas s ON s.schema_id = t.schema_id
JOIN sys.index_columns ic ON ic.object_id = i.object_id AND ic.index_id = i.index_id
JOIN sys.columns c        ON c.object_id  = ic.object_id AND c.column_id = ic.column_id
WHERE i.type > 0            -- exclude heaps
ORDER BY t.name, i.is_primary_key DESC, i.name, ic.key_ordinal;

---

5. Boarding-price types — ArtZustiegPreis

Answered 2026-08-04: this is a tariff-scheme lookup, not a direction lookup. Active schemes are 36 Woche26/27, 37 WE26/27, 38 Event26/27, 39 SER26/27 (plus legacy A–T at Status = 'L'). There is no Hinfahrt/Rückfahrt concept in this table.

SELECT IDArtZustiegPreis, Code, Kurztext, Langtext, Status, Hierarchie
FROM dbo.ArtZustiegPreis
ORDER BY Hierarchie, IDArtZustiegPreis;

6. Configured prices for boarding point 18

SELECT zp.IDZustiegPreis,
       zp.IDZustieg_FS,
       z.Zustiegcode, z.Ort, z.Plz,
       zp.IDArtZustiegPreis_FS,
       ap.Code AS PreisArtCode, ap.Kurztext AS PreisArt,
       zp.Preis, zp.PreisProvisionPflichtig,
       zp.Sockelbetrag, zp.SockelbetragProvisionPflichtig
FROM dbo.ZustiegPreis zp
JOIN dbo.Zustieg z            ON z.IDZustieg = zp.IDZustieg_FS
LEFT JOIN dbo.ArtZustiegPreis ap ON ap.IDArtZustiegPreis = zp.IDArtZustiegPreis_FS
WHERE zp.IDZustieg_FS = 18
ORDER BY zp.IDArtZustiegPreis_FS;

Answered 2026-08-04: three rows for point 18 (ESS-Route, Essen), one per active tariff scheme — 36 → 17,90, 37 → 12,90, 38 → 19,90. Prices are keyed by (point, tariff scheme), not by direction. The export's <zustiege_rueck preis="17,90"> is therefore the same row surfaced through the return service, not a separately configured return price.

7. How the two bus services resolve their boarding price — the mechanism

LeistungBefoerderung carries IDArtZustiegPreis_FS and Richtung_Hin_Rueck, so each transport service selects which price type applies. This query shows what outbound service 194779 and return service 194776 are configured to use.

SELECT l.IDLeistung, l.Bezeichnung, l.Art, l.Unterart,
       lb.Richtung_Hin_Rueck,
       lb.IDArtZustiegPreis_FS,
       ap.Code AS PreisArtCode, ap.Kurztext AS PreisArt,
       lb.IDArtZustiegZeitplan_FS,
       lb.ZustiegeGesperrt,
       lb.IDZustieg_FS_Basisstelle,
       lb.ZustiegIDs
FROM dbo.LeistungBefoerderung lb
JOIN dbo.Leistung l              ON l.IDLeistung = lb.IDLeistung_FS
LEFT JOIN dbo.ArtZustiegPreis ap ON ap.IDArtZustiegPreis = lb.IDArtZustiegPreis_FS
WHERE l.IDLeistung IN (194779, 194776);

Answered 2026-08-04 — configuration is symmetric, config cause excluded:

Leistung	Richtung	IDArtZustiegPreis_FS	ZustiegeGesperrt	ZustiegIDs
194779 Bus-Hinfahrt	H	36 (Woche26/27)	0	3,4,18,86,320,470,473
194776 Bus-Rückfahrt	R	36 (Woche26/27)	0	3,4,18,86,320,470,473

Both legs resolve to the same 17,90 and both have point 18 enabled. The differing BusPro behaviour is therefore not explained by product configuration.

8. Surcharge types — ArtZuschlag

Answered 2026-08-04: a discounts/fees lookup (Zwei-Wochen-Rabatt, Kinderermäßigung, Bearbeitungsgebühr, Preiskorrektur, …). No direction concept and no Abfahrtsortzuschlag entry — so the SON/TRA positions in the booking response are computed transport surcharges, not BuchungTNZuschlag rows. Query 9 below is aimed at the wrong table and can be skipped.

SELECT IDArtZuschlag, Code, Kurztext, Langtext, Status, Hierarchie
FROM dbo.ArtZuschlag
ORDER BY Hierarchie, IDArtZuschlag;

9. Booked surcharge rows for an affected booking

The decisive one if a completed booking exhibiting the fault is available: does the working booking persist two BuchungTNZuschlag rows and the failing one only one?

SELECT b.Vorgang,
       tn.IDBuchungTN,
       z.IDBuchungTNZuschlag,
       z.BZText, z.Text,
       z.Art,
       z.IDZustieg_FS,
       az.Code AS ZuschlagArtCode, az.Kurztext AS ZuschlagArt,
       z.BZPreisMitProvision, z.BZPreisOhneProvision
FROM dbo.BuchungTNZuschlag z
JOIN dbo.BuchungTN tn     ON tn.IDBuchungTN = z.IDBuchungTN_FS
JOIN dbo.Buchung b        ON b.IDBuchung    = tn.IDBuchung_FS
LEFT JOIN dbo.ArtZuschlag az ON az.IDArtZuschlag = z.IDArtZuschlag_FS
WHERE b.Vorgang IN ('<Vorgang-1>', '<Vorgang-2>')
ORDER BY b.Vorgang, tn.IDBuchungTN, z.IDBuchungTNZuschlag;

Caveat: both captured cases were <buchungsart>Anfrage</buchungsart> inquiries, and our flow only commits at step 4 — so these rows may not exist for them. Use the Vorgang of an actually completed booking that showed the short total. If none exists, skip this query; queries 5–7 carry the argument on their own.

---

What to send back

Query 1's result, plus 5, 6, 7 and 8 as plain result sets — those four are small and answer the pricing question directly. Queries 2–4 only if query 1 shows drift, as CSV files replacing the ones in this directory.