June 16, 2016 at 4:37 am
Eric M Russell (6/15/2016)
Brandie Tarvin (6/10/2016)
Steve Thompson-454462 (6/10/2016)
Brandie Tarvin (6/9/2016)
Y.B. (6/9/2016)
So what do you say when your boss announces in a meeting that 'we' should look into NoSQL? No really...what do you say?I realize that there are good use cases for both NoSQL and traditional RDBMS but when your whole knowledge base is in one area, it kind of feels like a gut punch. Oh well, I guess I ought to spin up a Linux instance on Azure and start playing (learning).
That is a good question. Where does one even start to look into NoSQL?
Again, seriously. Because I've pretty much ignored all the NoSQL stuff on account of not currently using it at work.
One way to explore this is to look at data sets that aren't necessarily relational. For example, event logs can be injected into Mongo pretty easily. Any data where you don't necessarily have a fixed schema on write is a good candidate; so maybe external data sources that you need to on-board.
Need to persisting payloads from API calls but parsing them on write to map them to your relational model is difficult/expensive? Again, Mongo is perfect for this.
If you're using a SQL Server column to store XML or (especially) JSON you may want to look into a non-relational document store (though it looks like 2016 has some nice JSON functionality).
Another example, might be using ElasticSearch for fuzzy/full-text searching (ES is really a document store under the hood).
If these data sets have keys that reference objects in your relational model, the data "joining" would be handled on read - so you need a middle tier that can handle it.
One plus for most NoSQL systems is that they are generally easier to scale horizontally than MS SQL, so if you have a system that is heavy on reads and needs scalability (e.g. a search engine), modeling that in something non-relational might be a win.
My org is keeping an open mind about NoSQL though all of our core systems remain firmly relational. We have committed heavily to ElasticSearch as our search engine.
Thanks for the reply.
I wonder... Our reporting team has a database where they essentially flatten all our data into giant tables so their reports read quicker. Would this be a candidate for a NoSQL solution?
Clustered ColumnStore (available in SQL Server 2014+) is good for very large flat tables. For denormalized tables with lots of repeatative text columns, I've seen it reduce both disk storage and query runtime by 90%. Your reporting team dump one of their full blown reports into a staging table and then create a clustered columnstore index on top of it. You'll see what I'm talking about.
Would LOVE to do that but we're just now getting on SQL 2012. It'll be forever before we get up to something using that.
June 16, 2016 at 4:44 am
Forgive me, but after everything going on this week, I feel the need to be political.
SET PSA = ON;
Imagine if you will, homeless data wandering the internet. Orphaned bits and bytes, seeking connection and a home. Imagine if you will, how cold, how lonely they are. Always looking for attention, for love.
Is it any wonder, then, that they resort to fanciful, crazy ploys? Is it any wonder they fall into depravity, becoming corrupt and corrupting those around them?
Stop Database Corruption now. Join your local SQL Server User Group and learn how to help prevent the damage done to these innocent data. It's everyone's responsibility.
SET PSA = OFF;
EDIT: It's not data that's bad. It's data entry clerks. It's the users that enter the data that's bad.
June 16, 2016 at 6:05 am
ChrisM@Work (6/16/2016)
jasona.work (6/15/2016)
/me jiggles door handle.../me knocks softly, tries to open door...
/me pokes head into room, quietly calls out "hello? Anybody here?"
/me leaves when no one answers...
Is this you testing the water having got home late and drunk, Jason?
Iahm nah drunk!
youre drunk!!
:hehe:
Nah, just felt it was too quiet around here...
June 16, 2016 at 6:11 am
Everyone say goodbye to Texas. I do believe it's going to disappear soon. :w00t:
On a serious note, I'm really hoping no one gets hurt when this happens.
June 16, 2016 at 6:46 am
I had the privilege of meeting Grant last night at the local PASS meeting. Liked his talk. Gave me a few (many) things to check today.
June 16, 2016 at 7:24 am
djj (6/16/2016)
I had the privilege of meeting Grant last night at the local PASS meeting. Liked his talk. Gave me a few (many) things to check today.
WHOOP!
It was nice to meet you as well. Love my Thread Brethren.
"The credit belongs to the man who is actually in the arena, whose face is marred by dust and sweat and blood"
- Theodore Roosevelt
Author of:
SQL Server Execution Plans
SQL Server Query Performance Tuning
June 16, 2016 at 7:36 am
Brandie Tarvin (6/16/2016)
Everyone say goodbye to Texas. I do believe it's going to disappear soon. :w00t:On a serious note, I'm really hoping no one gets hurt when this happens.
That's what happens for taking Mexico away from the mexicans. Just kidding, Texas wanted to be separated from Mexico anyway.
June 16, 2016 at 7:46 am
Luis Cazares (6/16/2016)
Brandie Tarvin (6/16/2016)
Everyone say goodbye to Texas. I do believe it's going to disappear soon. :w00t:On a serious note, I'm really hoping no one gets hurt when this happens.
That's what happens for taking Mexico away from the mexicans. Just kidding, Texas wanted to be separated from Mexico anyway.
That was a long time ago. Today it seems Texas would rather be separated from the U.S. 😀
_______________________________________________________________
Need help? Help us help you.
Read the article at http://www.sqlservercentral.com/articles/Best+Practices/61537/ for best practices on asking questions.
Need to split a string? Try Jeff Modens splitter http://www.sqlservercentral.com/articles/Tally+Table/72993/.
Cross Tabs and Pivots, Part 1 – Converting Rows to Columns - http://www.sqlservercentral.com/articles/T-SQL/63681/
Cross Tabs and Pivots, Part 2 - Dynamic Cross Tabs - http://www.sqlservercentral.com/articles/Crosstab/65048/
Understanding and Using APPLY (Part 1) - http://www.sqlservercentral.com/articles/APPLY/69953/
Understanding and Using APPLY (Part 2) - http://www.sqlservercentral.com/articles/APPLY/69954/
June 16, 2016 at 7:52 am
Brandie Tarvin (6/16/2016)
Eric M Russell (6/15/2016)
Brandie Tarvin (6/10/2016)
Steve Thompson-454462 (6/10/2016)
Brandie Tarvin (6/9/2016)
Y.B. (6/9/2016)
So what do you say when your boss announces in a meeting that 'we' should look into NoSQL? No really...what do you say?I realize that there are good use cases for both NoSQL and traditional RDBMS but when your whole knowledge base is in one area, it kind of feels like a gut punch. Oh well, I guess I ought to spin up a Linux instance on Azure and start playing (learning).
That is a good question. Where does one even start to look into NoSQL?
Again, seriously. Because I've pretty much ignored all the NoSQL stuff on account of not currently using it at work.
One way to explore this is to look at data sets that aren't necessarily relational. For example, event logs can be injected into Mongo pretty easily. Any data where you don't necessarily have a fixed schema on write is a good candidate; so maybe external data sources that you need to on-board.
Need to persisting payloads from API calls but parsing them on write to map them to your relational model is difficult/expensive? Again, Mongo is perfect for this.
If you're using a SQL Server column to store XML or (especially) JSON you may want to look into a non-relational document store (though it looks like 2016 has some nice JSON functionality).
Another example, might be using ElasticSearch for fuzzy/full-text searching (ES is really a document store under the hood).
If these data sets have keys that reference objects in your relational model, the data "joining" would be handled on read - so you need a middle tier that can handle it.
One plus for most NoSQL systems is that they are generally easier to scale horizontally than MS SQL, so if you have a system that is heavy on reads and needs scalability (e.g. a search engine), modeling that in something non-relational might be a win.
My org is keeping an open mind about NoSQL though all of our core systems remain firmly relational. We have committed heavily to ElasticSearch as our search engine.
Thanks for the reply.
I wonder... Our reporting team has a database where they essentially flatten all our data into giant tables so their reports read quicker. Would this be a candidate for a NoSQL solution?
Clustered ColumnStore (available in SQL Server 2014+) is good for very large flat tables. For denormalized tables with lots of repeatative text columns, I've seen it reduce both disk storage and query runtime by 90%. Your reporting team dump one of their full blown reports into a staging table and then create a clustered columnstore index on top of it. You'll see what I'm talking about.
Would LOVE to do that but we're just now getting on SQL 2012. It'll be forever before we get up to something using that.
Aye, because of not being able to quickly jump to 2014, NoSQL becomes an option on existing hardware. It's free, MPP and can be installed right away. SQL Server cannot unless you shell out the $$$.
I think good candidates for NoSQL involve large datasets that are not easily handled in traditional RDBMS. If not easily, maybe also data that is not really ready to be in the traditional RDBMS as data can be expensive in RDBMS versus NoSQL. Hence the marketing terms, "Data Lake". Therefore, having that cluster of raw data feeding your RDBMS is all the rave and a good idea if you see the value.
For example, I work with a lot of highly relational log data. It never changes. It can go right into a RDBMS. The issue is the size of the logs causes the RDBMS to crawl due to the business requirements sometimes. Augmenting that to NoSQL and creating a pipeline between NoSQL and SQL Server allows SQL Server to focus on the end result that eventually feeds into reporting (Data Lake -> Data Warehouse -> Data Mart -> Report).
June 16, 2016 at 8:29 am
Sean Lange (6/16/2016)
Luis Cazares (6/16/2016)
Brandie Tarvin (6/16/2016)
Everyone say goodbye to Texas. I do believe it's going to disappear soon. :w00t:On a serious note, I'm really hoping no one gets hurt when this happens.
That's what happens for taking Mexico away from the mexicans. Just kidding, Texas wanted to be separated from Mexico anyway.
That was a long time ago. Today it seems Texas would rather be separated from the U.S. 😀
Actually, I think Texas wants to become it's own country, separate from everyone. 😛
June 16, 2016 at 9:01 am
Sean Lange (6/16/2016)
Luis Cazares (6/16/2016)
Brandie Tarvin (6/16/2016)
Everyone say goodbye to Texas. I do believe it's going to disappear soon. :w00t:On a serious note, I'm really hoping no one gets hurt when this happens.
That's what happens for taking Mexico away from the mexicans. Just kidding, Texas wanted to be separated from Mexico anyway.
That was a long time ago. Today it seems Texas would rather be separated from the U.S. 😀
Maybe Texans want to live in the new Atlantis.
June 16, 2016 at 2:26 pm
I'm sorry, this post infuriates me.
Integer division was taught within the first three weeks of my first programming class in my first semester of school.
Michael L John
If you assassinate a DBA, would you pull a trigger?
To properly post on a forum:
http://www.sqlservercentral.com/articles/61537/
June 16, 2016 at 2:48 pm
Michael L John (6/16/2016)
I'm sorry, this post infuriates me.Integer division was taught within the first three weeks of my first programming class in my first semester of school.
Actually, I remember learning about integer arithmetic in the first grade. For example, 1 divided by 2 equals 0 remainder 1.
June 16, 2016 at 2:58 pm
Lynn Pettis (6/16/2016)
Michael L John (6/16/2016)
I'm sorry, this post infuriates me.Integer division was taught within the first three weeks of my first programming class in my first semester of school.
Actually, I remember learning about integer arithmetic in the first grade. For example, 1 divided by 2 equals 0 remainder 1.
Is that old math, new math, or common core math???
Michael L John
If you assassinate a DBA, would you pull a trigger?
To properly post on a forum:
http://www.sqlservercentral.com/articles/61537/
Viewing 15 posts - 54,601 through 54,615 (of 66,688 total)
You must be logged in to reply to this topic. Login to reply