I have a Databricks delta table of financial transactions that is essentially a running log of all changes that ever took place on each record. Each record is uniquely identified by 3 keys. So given that uniqueness, each record can have multiple instances in this table. Each representing a historical entry of a change(across one or more columns of that record) Now if I wanted to find out cases where a specific column value changed I can easily achieve that by doing something like this -->
SELECT t1.Key1, t1.Key2, t1.Key3, t1.Col12 as "Before", t2.Col12 as "After"
from table1 t1 inner join table t2 on t1.Key1= t2.Key1 and t1.Key2 = t2.Key2
and t1.Key3 = t2.Key3 where t1.Col12 != t2.Col12
However, these tables have a large amount of columns. What I'm trying to achieve is a way to identify any columns that changed in a self-join like this. Essentially a list of all columns that changed. I don't care about the actual value that changed. Just a list of column names that changed across all records. Doesn't even have to be per row. But the 3 keys will always be excluded, since they uniquely define a record.
Essentially I'm trying to find any columns that are susceptible to change. So that I can focus on them dedicatedly for some other purpose.
Any suggestions would be really appreciated.
Databricks has change data feed (CDF / CDC) functionality that can simplify these type of use cases. https://docs.databricks.com/delta/delta-change-data-feed.html
i'm currently trying to optimize some kind of query of 2 rather large tables, which are characterized like this:
Table 1: id column - alphanumerical, about 300mil unique ids, more than 1bil rows overall
Table 2: id column - identical semantics, about 200mil unique ids, more than 1bil rows overall
Lets say on a given day, 17.03. i want to join those two tables on id.
Table 1 is left, table 2 is right, i get like 90% of matches, meaning table 2 has like 90% of those ids present in table 1.
One week later, said table 1 did not change (could but to make explanation easier, consider it didn't), table 2 was updated and now contains more records. I do the join again and now, from the former missing ids some came up, so i got like 95% matches now.
In general, table1.id has some matches with table2.id at a given time which might change on a day-per-day base.
I now want to optimize this join and came up on the bucketing feature. Is this possible?
Example:
1st join: id "ABC123" is present in table1, not in table2. ABC123 gets sorted into a certain bucket, e.g. "1".
2nd join (week later): id "ABC123" now came up in table2; how can it be ensured it comes into the bucket on table 2 which then is co-located with table 1?
Or am i having a general problem of understanding how it works?
I am currently trying to dig into Cassandra's data model and its relation to Bigtable, but ended up with a strong headache concerning the Column Family concept.
Mainly my question was asked and already answered. However, I'm not satisfied with the answers :)
Firstly I've read the Bigtable paper especially concerning its data model, i.e. how data is stored. As far as I understood each table in Bigtable basically relies on a multi-dimensional sparse map with the dimensions row, column and time. The map is sorted by rows. Columns can be grouped with the name convention family:qualifier to a column family. Therefore, a single row can contain multiple column families (see the example figure in the paper).
Although it is stated that Cassandra relies on Bigtable data model, I read multiple times that in Cassandra a column family contains multiple rows and is to some extent comparable to a table in relational data stores. Isn't this contrary to Bigtable's approach, where a row could contain multiple column families? What comes first, the column family or row :)? Are these concepts even comparable?
The answer you linked to was from 6 years ago, and a lot has changed in Cassandra since. When Cassandra started out, its data model was indeed based on BigTable's. A row of data could include any number of columns, each of these columns has a name and a value. A row could have a thousand different columns, and a different row could have a thousand other columns - rows do not have to have the same columns. Such a database is called "schema-less", because there is no schema that each row needs to adhere to.
But Toto, we're not in Kansas any more - and Cassandra's model changed in focus (though not in essense) since, and I'll try to explain how and why:
As Cassandra matured, its developers started to realize that schema-less isn't as great as they once thought it was. Schemas are valuable in ensuring application correctness. Moreover, one doesn't normally get to 1000 columns in a single row just because there are 1000 individually-named fields in one record. Rather, the more common case is that the record actually contains 200 entries, each with 5 fields. The schema should fix these 5 fields that every one of these entries should have, and what defines each of these separate entries is called a "clustering key". So around the time of Cassandra 0.8, six years ago, these ideas where introduced to Cassandra as the "CQL" (Cassandra Query Language).
For example, in CQL one declares that a column-family (which was dutifully renamed "table") has a schema, with a known list of fields:
CREATE TABLE groups (
groupname text,
username text,
email text,
age int,
PRIMARY KEY (groupname, username)
)
This schema says that each wide row in the table (now, in modern Cassandra, this was renamed a "partition") with the key "groupname" is a a possibly long list of users, each with username, email and age fields. The first name in the "PRIMARY KEY" specifier is the partition key (it determines the key of the wide rows), and the second is called the clustering key (it determines the key of the small rows that together make up the wide rows).
Despite the new CQL dressup, Cassandra continued to implement these new concepts using the good-old-BigTable-wide-row-without-schema implementation. For example, consider that our data has a group "mygroup" with two people, (john, john#somewhere.com, 27) and (joe, joe#somewhere.com, 38). Cassandra adds the following four column names->values to the wide row:
john:email -> john#somewhere.com
john:age -> 27
joe:email -> joe#somewhere.com
joe:age -> 27
Note how we ended up with a wide row with 4 columns - 2 non-key fields per row (email and age), multiplied by the number of rows in the partition (2). The clustering key field "username" no longer appears anywhere as the value, but rather as part of the column's name! So If we have two username values "john" and "joe", We have some columns prefixed "john" and some columns prefixed "joe", and when we read the column "joe:email" we know this is the value of the email field of the row which has username=joe.
Cassandra still has this internal duality - converting the user-facing CQL rows and clustering keys into old-style wide rows. Until recently, Cassandra's on-disk format known as "SSTables" was still schema-less and used composite names as shown above for column names. I wrote a detailed description of the SSTable format on Scylla's site https://github.com/scylladb/scylla/wiki/SSTables-Data-File (Scylla is a more efficient C++ re-implementation of Cassandra to which I contribute). However, column names are very inefficient in this format so Cassandra recently (in version 3.0) switched to a different file format, which for the first time, accepts clustering keys and schema-full rows as first class citizens. This was the last nail in the coffin of the schema-less Cassandra from 7 years ago. Cassandra is now schema-full, all the way.
I'm so confused.
When to use them and how to determine which one to use?
If a column is index/primary key/row key, could it be duplicated?
I want to create a column family to store some many-to-many info, for example, one column is the given name and the other is surname. One given name can related to many surnames, and one surname could have different given names.
I need to query surnames by a given name, and the given names by a specified surname too.
How to create the table?
Thanks!
Cassandra is a NoSQL database, and as such has no such concept of many-to-many relationships. Ideally a table should not have anything other than a primary key. In your case the right way to model it in Cassandra is to create two tables, one with name as the primary key and the other with surname as the primary key
When you need to query by either key, you need to query the table that has that key as the primary key
EDIT:
From the Cassandra docs:
Cassandra's built-in indexes are best on a table having many rows that
contain the indexed value. The more unique values that exist in a
particular column, the more overhead you will have, on average, to
query and maintain the index. For example, suppose you had a races
table with a billion entries for cyclists in hundreds of races and
wanted to look up rank by the cyclist. Many cyclists' ranks will share
the same column value for race year. The race_year column is a good
candidate for an index.
Do not use an index in these situations:
On high-cardinality columns for a query of a huge volume of records for a small number of results.
In tables that use a counter column On a frequently updated or deleted column.
To look for a row in a large partition unless narrowly queried.
Reading several papers and documents on internet, I found many contradictory information about the Cassandra data model. There are many which identify it as a column oriented database, other as a row-oriented and then who define it as a hybrid way of both.
According to what I know about how Cassandra stores file, it uses the *-Index.db file to access at the right position of the *-Data.db file where it is stored the bloom filter, column index and then the columns of the required row.
In my opinion, this is strictly row-oriented. Is there something I'm missing?
If you take a look at the Readme file at Apache Cassandra git repo, it says that,
Cassandra is a partitioned row store. Rows are organized into tables
with a required primary key.
Partitioning means that Cassandra can distribute your data across
multiple machines in an application-transparent matter. Cassandra will
automatically repartition as machines are added and removed from the
cluster.
Row store means that like relational databases, Cassandra organizes
data by rows and columns.
Column oriented or columnar databases are stored on disk column wise.
e.g: Table Bonuses table
ID Last First Bonus
1 Doe John 8000
2 Smith Jane 4000
3 Beck Sam 1000
In a row-oriented database management system, the data would be stored like this: 1,Doe,John,8000;2,Smith,Jane,4000;3,Beck,Sam,1000;
In a column-oriented database management system, the data would be stored like this:
1,2,3;Doe,Smith,Beck;John,Jane,Sam;8000,4000,1000;
Cassandra is basically a column-family store
Cassandra would store the above data as,
"Bonuses" : {
row1 : { "ID":1, "Last":"Doe", "First":"John", "Bonus":8000},
row2 : { "ID":2, "Last":"Smith", "First":"Jane", "Bonus":4000}
...
}
Also, the number of columns in each row doesn't have to be the same. One row can have 100 columns and the next row can have only 1 column.
Read this for more details.
Yes, the "column-oriented" terminology is a bit confusing.
The model in Cassandra is that rows contain columns. To access the smallest unit of data (a column) you have to specify first the row name (key), then the column name.
So in a columnfamily called Fruit you could have a structure like the following example (with 2 rows), where the fruit types are the row keys, and the columns each have a name and value.
apple -> colour weight price variety
"red" 100 40 "Cox"
orange -> colour weight price origin
"orange" 120 50 "Spain"
One difference from a table-based relational database is that one can omit columns (orange has no variety), or add arbitrary columns (orange has origin) at any time. You can still imagine the data above as a table, albeit a sparse one where many values might be empty.
However, a "column-oriented" model can also be used for lists and time series, where every column name is unique (and here we have just one row, but we could have thousands or millions of columns):
temperature -> 2012-09-01 2012-09-02 2012-09-03 ...
40 41 39 ...
which is quite different from a relational model, where one would have to model the entries of a time series as rows not columns. This type of usage is often referred to as "wide rows".
You both make good points and it can be confusing. In the example where
apple -> colour weight price variety
"red" 100 40 "Cox"
apple is the key value and the column is the data, which contains all 4 data items. From what was described it sounds like all 4 data items are stored together as a single object then parsed by the application to pull just the value required. Therefore from an IO perspective I need to read the entire object. IMHO this is inherently row (or object) based not column based.
Column based storage became popular for warehousing, because it offers extreme compression and reduced IO for full table scans (DW) but at the cost of increased IO for OLTP when you needed to pull every column (select *). Most queries don't need every column and due to compression the IO can be greatly reduced for full table scans for just a few columns. Let me provide an example
apple -> colour weight price variety
"red" 100 40 "Cox"
grape -> colour weight price variety
"red" 100 40 "Cox"
We have two different fruits, but both have a colour = red. If we store colour in a separate disk page (block) from weight, price and variety so the only thing stored is colour, then when we compress the page we can achieve extreme compression due to a lot of de-duplication. Instead of storing 100 rows (hypothetically) in a page, we can store 10,000 colour's. Now to read everything with the colour red it might be 1 IO instead of thousands of IO's which is really good for warehousing and analytics, but bad for OLTP if I need to update the entire row since the row might have hundreds of columns and a single update (or insert) could require hundreds of IO's.
Unless I'm missing something I wouldn't call this columnar based, I'd call it object based. It's still not clear on how objects are arranged on disk. Are multiple objects placed into the same disk page? Is there any way of ensuring objects with the same meta data go together? To the point that one fruit might contain different data than another fruit since its just meta data or xml or whatever you want to store in the object itself, is there a way to ensure certain matching fruit types are stored together to increase efficiency?
Larry
The most unambiguous term I have come across is wide-column store.
It is a kind of two-dimensional key-value store, where you use a row key and a column key to access data.
The main difference between this model and the relational ones (both row-oriented and column-oriented) is that the column information is part of the data.
This implies data can be sparse. That means different rows don't need to share the same column names nor number of columns. This enables semi-structured data or schema free tables.
You can think of wide-column stores as tables that can hold an unlimited number of columns, and thus are wide.
Here's a couple of links to back this up:
This mongodb article
This Datastax article mentions it too, although it classifies Cassandra as a key-value store.
This db-engines article
This 2013 article
Wikipedia
Column Family does not mean it is column-oriented. Cassandra is column family but not column-oriented. It stores the row with all its column families together.
Hbase is column family as well as stores column families in column-oriented fashion. Different column families are stored separately in a node or they can even reside in different node.
IMO that's the wrong term used for Cassandra. Instead, it is more appropriate to call it row-partition store. Let me provide you some details on it:
Primary Key, Partitioning Key, Clustering Columns, and Data Columns:
Every table must have a primary key with unique constraint.
Primary Key = Partition key + Clustering Columns
# Example
Primary Key: ((col1, col2), col3, col4) # primary key uniquely identifies a row
# we need to choose its components partition key
# and clustering columns so that each row can be
# uniquely identified
Partition Key: (col1, col2) # decides on which node to store the data
# partitioning key is mandatory, and it
# can be made up of one column or multiple
Clustering Columns: col3, col4 # decides arrangement within a partition
# clustering columns are optional
Partition key is the first component of Primary key. Its hashed value is used to determine the node to store the data. The partition key can be a compound key consisting of multiple columns. We want almost equal spreads of data, and we keep this in mind while choosing primary key.
Any fields listed after the Partition Key in Primary Key are called Clustering Columns. These store data in ascending order within the partition. The clustering column component also helps in making sure the primary key of each row is unique.
You can use as many clustering columns as you would like. You cannot use the clustering columns out of order in the SELECT statement. You may choose to omit using a clustering column in you SELECT statement. That's OK. Just remember to sue them in order when you are using the SELECT statement. But note that, in your CQL query, you can not try to access a column or a clustering column if you have not used the other defined clustering columns. For example, if primary key is (year, artist_name, album_name) and you want to use city column in your query's WHERE clause, then you can use it only if your WHERE clause makes use of all of the columns which are part of primary key.
Tokens:
Cassandra uses tokens to determine which node holds what data. A token is a 64-bit integer, and Cassandra assigns ranges of these tokens to nodes so that each possible token is owned by a node. Adding more nodes to the cluster or removing old ones leads to redistributing these token among nodes.
A row's partition key is used to calculate a token using a given partitioner (a hash function for computing the token of a partition key) to determine which node owns that row.
Cassandra is Row-partition store:
Row is the smallest unit that stores related data in Cassandra.
Don't think of Cassandra's column family (that is, table) as a RDBMS table, but think of it as a dict of a dict (here dict is data structure similar to Python's OrderedDict):
the outer dict is keyed by a row key (primary key): this determines which partition and which row in partition
the inner dict is keyed by a column key (data columns): this is data in dict with column names as keys
both dict are ordered (by key) and are sorted: the outer dict is sorted by primary key
This model allows you to omit columns or add arbitrary columns at any time, as it allows you to have different data columns for different rows.
Cassandra has a concept of column families(table), which originally comes from BigTable. Though, it is really misleading to call them column-oriented as you mentioned. Within each column family, they store all columns from a row together, along with a row key, and they do not use column compression. Thus, the Bigtable model is still mostly row-oriented.