dev-resources.site
for different kinds of informations.
Working with NULL in Databases. Turn Your Frustration Into Delight
NULL values are one of the most frequent frenemies of almost all database administrators. For many, theyâre simply a symbol that no value exists in a column â for others, theyâre a symbol of problems within their database instances. Join us as we explore them and tell you how best to act in specific scenarios.
NULL values are some of the most popular values within database management systems: thereâs no need to be a database expert to tell that NULL means nothing, and thatâs exactly what NULL signifies â itâs a symbol to signify that no values exist in a column. Thatâs not to say that NULL cannot be problematic though.
The Basics of NULL
Many database administrators set NULL as the default value of their columns when creating a table by running a query like so:
Now when any data is inserted into the column named column_1, the default â âpre-builtâ if you will â value will be NULL. Everything can also be done the other way around too, though â columns can be specified to not accept NULL values and have a default empty value like so:
Now, when anything is inserted into the table that is a NULL value, the RDBMS will respond with an error that looks like the following â it will inform us that a column cannot be NULL:
But thatâs the least of our problems... Keep reading â weâll tell you why in a second.
Problems with NULL
To start with, the most frequent problem related to NULL values is related to the confusion of NULL values and empty strings (rows with no value.) Many newcomers to the SQL world think these two things are the same when in reality, theyâre not. Take these two queries for an example:
The first query will insert an empty string into the email column, while the second query will insert NULL. Now look at the output of this query:
The query above probably looks a little different than the queries youâre used to, doesnât it? The WHERE clause doesnât specify an equal (=) sign, but instead, it specifies IS NULL. Letâs try to do everything the ordinary way:
âEmpty result set fetched.â Why? Because something = nothing (NULL) is never true.
To test for null, we need to use the IS [NOT] NULL clause rather than search for a match â if we need a function, we should probably turn to IFNULL() instead. In the previous example, weâve successfully confused MySQLâ are we searching for an email with a value of NULL? Or are we searching for an email without a value, but with a default value of NULL? The question is not clear â so is the answer.
NULL also has a couple of things unique to itself â when using some data types (weâre talking about integers and floating-point values), if the column is specified to increment automatically, the next value is inserted into it. Look at the following example:
Technically, 3 NULL values were inserted, but as the column was set to increment automatically, the values were populated with numbers.
As far as indexes are concerned, columns having NULL values can easily be indexed and that isnât the cause of much problems, however, bear in mind that NULL values also occupy space within MySQL. Everything works like this:
When MyISAM is being used, a field set to NOT NULL will take up less space on the disk.
When MyISAM is in use, a field set to NULL as the default value, will take up more space on the disk.
NULL does take up storage space, but in order to see any real impact of storing NULL values within a database, you would literally have to have billions of such values (at that point NULL values are the least of your problems.) In other words, NULL values donât require much storage space in and of itself for many storage engines within MySQL, however, if the NDB storage engine (or NDBCluster) is in use, things change a little â it reserves 4 bytes per row for NULL values. However, the impact of this is probably negligible.
However, if we come to partitions, thatâs where NULL values cause the most problems for many advanced DBAs. We get into them below.
Partitioning and NULL
One more problem with NULL values that is specific to MySQL and its associates is partitioning. MySQL as an RDBMS doesnât exactly disallow or forbid the NULL value to be used, but it does treat the NULL value as being âless than any other value.â Hereâs how everything works when our data is partitioned:
If you insert NULL values into a partition on a table partitioned by RANGE, the row is inserted into the lowest partition possible.
If LIST partitioning is in use, NULL values can only be inserted into the specific partition if NULL is specified in the list of possible values. Otherwise, weâll get the error 1504: ERROR 1504 (HY000): Table has no partition for value NULL
If HASH or KEY partitioning is being used, NULL is treated the same as if the expression would return 0 (more information can be found over at the documentation.)
Summary
In this blog, weâve walked you through the concept of NULL values in relational database management systems, with the examples being based on MySQL. While the concept of NULL values might seem problematic and perhaps even frightening to some, it doesnât have to be that way â when properly managed, NULL values can become a really good friend of yours.
We hope that youâve enjoyed reading this blog, come back to the DbVisualizer blog in the future to learn even more about databases and how they work, and until next time!
About the author
Lukas Vileikis is an ethical hacker and a frequent conference speaker. He runs one of the biggest & fastest data breach search engines in the world - BreachDirectory.com, frequently speaks at conferences and blogs in multiple places including his blog over at lukasvileikis.com.
Featured ones: