Getting started

From nothing to a loaded dataset. Four steps, roughly twenty minutes.

1. Get a token

Ask the customer system's IdentityServer for a token using the client_credentials grant and the scope ticketinghub-api. No user, no browser, no redirect — this is a server-to-server credential, valid for one hour.

Detail, including the one requirement that catches people out, is in Authentication.

2. Pull a dataset, naming your columns

GET https://<your-api-host>/api/ticketinghub/odata/v1/Tickets?$select=TicketInternalId,FaireventName,TicketGrossPrice,CurrencyShort
Authorization: Bearer <token>

Name the columns you actually load. A full ticket row carries 342 columns and weighs about 10 KB; most warehouses want a fraction of that. Projecting is also the single biggest lever on speed.

3. Follow the next link until it stops

Each response carries a page of rows and an @odata.nextLink. Request that link, repeat, and when a response comes back without one you have the complete dataset.

Paging is keyset-based, so the last page costs what the first one did, and rows written while you read cannot be skipped or duplicated. Do not build your own paging with $skip.

4. Do the same for the other three, then model in your warehouse

Tickets is the wide, denormalised centre. TicketUsages adds the scans, Surveys the survey answers, and StatisticGroups the article-to-group relation that the ticket rows cannot express on their own.

Building the extract turns these four steps into a production-ready loader, with retry, incremental loading and scheduling.

The reading order that saves time