Katie prepared herself for 98-364 database administration fundamentals using Microsoft book (with the same title) and its rather surprising for me to see simple errors that have been made in the book so I decided to do quick review and list some of error I managed to find below:
Page 9 ("take note"): "A foreign key is also known as a self-reference."
Comments: I must admit I didn't know that.... I've heard about self-referencing tables but not that a foreign key is also known as a self-reference? Can someone confirm that? Quick Google search doesn't seem to provide any evidence on that.
Page 13 (last section): "(...) again, the six main DDL statements are as follows: Use, Create, alter, drop, truncate, delete"
Comments: USE & Delete are NOT DDL statements!!
Page 16 ("take note"): "If you are deleting data from tables in a large database, use TRUNCATEl it is more efficient. USE DELETE for smaller databases.
Comments: Why do they use word "database"? It doesn't make sense? but coming to the point...... use truncate on big tables? hmmm hopefully not on production environment? because it wouldn't log deletes and it would fail if there are foreign key constraints and you require alter permission to do that. I use TRUNCATE only for clearing stage tables in data warehouse and is special cases (which is not often) as I have Foreign Keys that don't allow me to do that (Yes I do have FKs in a data warehouse! ETL can make a mistake as well :p)
Page 18 - Multiple Choice:
Question 4: Which of the following is not a DML statement?
Comments: Let me test my knowledge. I've never heard about Remove so I would say A... but wait! I know TRUNCATE is not DML; it is DDL! so D as well.... so question should be which of the following ARE not a DML statement (choose two).
Page 23: Approximate Numerics are: datetime? smalldatetime? and so on..
Comments: This is a human error and heading shouldn't say "approximate numerics" but "data and time" data types
Page 28 ("take note"): Use nchar when the sizes of the columns data entries will be similar. Use nvarchar when the sizes of the column data entries will vary considerably, for such things as binary files, image files, SQL Variant and UUID.
Comments: ok I think I'm completely confused with the last part. Can someone explain this to me?
Page 71 ("take note"): "Every table must have a primary key; without a primary key; it's not a valid table. By definition, a primary key must be unique and must have a value that is not null."
Comments: Yes.....if you mean logical primary key. You don't need physical primary key in star schema for fact tables and they are still valid tables! (I think)
Page 81 (question 2): Which of the following can help speed data retrieval:
a) a default constraint
b) a primary key constraint
c) a clustered index
d) a foreign key constraint
Comments: C sounds like a correct answer but wait! If I create a primary key constraint it will automatically create a clustered index (or non-clustered if I select correct option) and that should also speed up data retrieval! Can foreign key constraint speed up data retrieval? I don't know but I would imagines it might a little bit? Does anyone knows if FK can speed up data retrieval?
Page 82 (Question 8): How many clustered indexes can you have for a database?:
Comments: Someone is obsessed with a word "database". You can have as many clustered indexes in a database as many tables you have!
I finished! I skipped lesson 5 and bits in the middle as it is getting late!
Final words: Microsoft has created a very good book; with a few mistakes but still very good so I can honestly recommend it to everyone (after reading exceptions) that wants to dive into the world of databases!