SQL Humility

I answer the same question every time. I tried to give a number reflecting my language knowledge and ability to answer trivia to the interviewer and another number representing my confidence in that language. The second number I tried to emphasize.

"I am very comfortable and confident about SQL - even if I do not know all the commands / methods / skills / little things"

I do not know the reason, but I was proud to write SQL at some stage of my career. I thought that I was riveting. Recent work has noticed that there are many things to learn.

4. Select convert (varchar, getdate (), 107) (wait, wait, etc. Can you format dates as different format strings?).

Therefore, I am reading a lot of articles at http://use-the-index-luke.com/. It makes me feel smart and improves my SQL skills.

SQL-centric ETLs are usually built using languages ​​such as SQL, Presto, Hive, and so on. ETL jobs are usually declaratively defined, and almost all are centered around SQL and tables. Since it is necessary to create UDF in different languages ​​(Java, Python, etc.), it may be troublesome to create. Therefore, testing may be more difficult. This paradigm is very popular among data scientists. When a data scientist builds an ETL pipeline with two paradigms, I prefer naturally SQL center ETL. In fact, I think as a new data scientist I can learn data engineering early while working with the SQL paradigm. why? Because learning SQL is much simpler than learning Java and Scala (unless you are familiar with them), and you are better off learning new concepts in DE with new domains based on new languages ​​than best practices of DE You can concentrate on learning.

Everyone knows SQL. It is not your own query language like Mongo. SQL is beyond the scope of the engineering department and is almost anywhere in the organization. In terms of organizational expansion, employing new people who already know SQL is easier than finding a person using MongoDB. So temporarily decided that the following database is SQL compatible. It is not clear how to meet the benchmark requirements. Postgres and MySQL are far from obstacles we set up. After that, as is often the case, I found "Analysis database" which is Google's correct search term. We found that there is an entire set of SQL databases built and optimized for aggregate queries. One of the leaders in this field is Vertica, which was acquired by HP several years ago. Internally, Vertica is a columnar database, but externally it has a complete SQL interface.