{"id":7262,"date":"2019-07-26T12:17:59","date_gmt":"2019-07-26T19:17:59","guid":{"rendered":"http:\/\/www.couchbase.com\/blog\/?p=7262"},"modified":"2025-06-13T17:22:00","modified_gmt":"2025-06-14T00:22:00","slug":"analytics-love-the-doc-model-youre-with","status":"publish","type":"post","link":"https:\/\/www.couchbase.com\/blog\/es\/analytics-love-the-doc-model-youre-with\/","title":{"rendered":"Anal\u00edtica - Ama el modelo de m\u00e9dico con el que est\u00e1s"},"content":{"rendered":"

In business applications, the data is often modeled for a large number of concurrent low latency queries. If you want to gain insight by looking at trends, however, you end up wishing you had an entirely different data model. The traditional means for squaring this circle had been to move, transform, and load the data elsewhere, but this introduces its own raft of problems, including unacceptable latency, multiple sources of truth, and a lot of expense.<\/p>\n

Couchbase customers know that the Analytics service provides an easy way to handle real-time analytical and trend reporting on the data they have in production right now. An example of this recently came up when we worked with a customer looking to identify high-end customer activity associated with a corporate partner loyalty program. The underlying document model was clearly designed with the interactive application, rather than reporting, in mind. (This is not at all uncommon, as you may know from painful experience.) Let\u2019s take a quick look at the problem and how we solved it.<\/p>\n

Example document<\/h4>\n

The document model in our case (supporting an online booking application) is comprised of four sections. The first section includes basic document and app identifiers. The second describes the booking information about an excursion. The third contains details on one or more itineraries associated with the booking, along with passenger requirements for one or more passengers. The final section describes the corporate loyalty programs to which each of the passengers might belong.<\/p>\n

\n
{\r\n  \"_type\": \"booking\",\r\n  \"_header\": {\r\n    \"created\": 1562888960,\r\n    \"source\": \"app\",\r\n    \"version\": \"v1.1\"\r\n  },\r\n  \"booking\": {\r\n    \"status\": \"BOOKED\",\r\n    \"bookingType\": \"agency\",\r\n    \"details\": {\r\n      \"agent\": \"FBL33\",\r\n      \"contact\": \"Arlene\",\r\n      \"seats\": 2,\r\n      \"excursion\": {\r\n        \"embarking\": 1562958000,\r\n        \"equipment\": \"123X\",\r\n        \"line\": \"SRF\",\r\n        \"fromStation\": {\r\n          \"code\": \"LAX\",\r\n          \"facilityType\": 1\r\n        },\r\n        \"toStation\": {\r\n          \"code\": \"SOL\",\r\n          \"facilityType\": 2\r\n        },\r\n        \"bookingAgency\": \"PC\",\r\n        \"agencyType\": \"3\"\r\n      }\r\n    }\r\n  },\r\n  \"itinerary\": [\r\n    {\r\n      \"daysOnboard\": 1,\r\n      \"passengers\": [\r\n        {\r\n          \"passengerNumber\": 1,\r\n          \"specialAccomodations\": false\r\n        },\r\n        {\r\n          \"passengerNumber\": 2,\r\n          \"specialAccomodations\": false\r\n        }\r\n      ],\r\n      \"itineraryType\": \"business\"\r\n    }\r\n  ],\r\n  \"passengerDetails\": [\r\n    {\r\n      \"loyaltyId\": \"aaaabbbbccccdddd\",\r\n      \"passId\": 1,\r\n      \"programType\": {\r\n        \"corporatePartner\": true,\r\n        \"partnerId\": 1\r\n      }\r\n    },\r\n    {\r\n      \"loyaltyId\": \"eeeeffffgggghhhh\",\r\n      \"passId\": 2,\r\n      \"programType\": {\r\n        \"corporatePartner\": false\r\n      }\r\n    }\r\n  ]\r\n}<\/pre>\n
 <\/p>\n<\/div>\n
Query elements<\/h4>\n
In order to complete the analysis, my customer was required to pull or filter on the following fields:<\/p>\n