Thanks for the awesome video! It would be great if you can create a detailed video with example of each type of dimension. This will help the people who are new to Data Modeling.
Hi Abhilash... 5:05, you mentioned 'invoice num' falls under dimensions not facts. That sounds confusing! invoice numbers are random and differs from transaction to transactions, an ideal fit for a fact table. How is it a dimension? Could you us get thru it
Yeah! Invoice Number is not a measure. Sum or any aggregate function won’t make sense on invoice number but usually on a measure like sales amount, you can apply aggregations. So invoice number is a dimension, but it’s a degenerate dimension and we keep it in fact table usually as I mentioned in the video.
Hey! We cannot say that because SCD Type 1 is a ‘dimension’, OLTP is a whole system design it can be in any level of normalization (Usually 3rd) and it is NOT restricted only to Dimensions.
@@AbhilashMarichi Thank you I'll try to be short. I wish to send customers a QR code as a receipt for a pack of minutes of time of a service (assisted stretch) with the goal of the QR code being able to record or refer to a data mart or small data warhouse where a updated balance can be maintained. For example, we offer 3 products. A pack 100 minutes, 200 minutes and 300 minutes. If a customer purchases a 100 minute pack, the QR code they are emailed as part of a receipt would then contain or record or refer to a record of their balance of 100 minutes. If they meet us to use the service, we wish to track the use of their minutes by scanning the QR code. A 1st scan would log the start of the service & a 2nd would mark the end, so if they had 100 minutes, and a1st scan of the QR code was done at 1pm and a 2nd at 130pm, 30 minutes of the service was used up, 30 would be deducted from the balance of 100 and the new balance of 70 minutes of service time remaining on their QR code invoice would remain. I wish to better understand options for how to organize this. Could times of a QR code scan be stored on one dimension table, the product (100, 200 or 300) on another, the customer info (address on another) and the credit card be the non volatile fact table at the center of a star schema? Or given the above can I get guidance on a general overview of how to organize the backend to store this info and reliably use QR codes this way? My question was longer to explain than I hoped but I hope I was clear. Thank you. Happy to share more information if you like.
You should look at building an OLTP system and NOT a data warehouse at this point for making your application to function optimally. I say this because, your writes to the database needs to be faster. Functionally separate the data into multiple tables and store it in 3rd Normal Form as much as possible. Once you have an operating OLTP then we can think of building a Data Warehouse.
Hey Rushikesh! I believe that is something related to Psychology! 😊 I am not sure if it’s related to Data Warehousing! Let me know if you find something related to DW!
@@AbhilashMarichi NO its not i was reading this book by Christopher named star schema where he have mention this 👉👉👉👉👉👉👉👉👉 " A very powerful analytic technique uses behavior patterns to analyze facts. Consider this question: “Are customers who generate over $1 million in sales receiving better discounts than those who generate $500,000 or less?” Orders are tracked by a fact. To answer this question, we must group the customers based on their past order activity. The groupings are then used as a dimension, and they are used to study discount A dimension that is computed based on facts is called a behavioral dimension"
Though we have 19 different Genders now. I would still say it is static from Dimensional modeling perspective and it allows the association between Person Dimension and Gender Dimension to change as needed.