BITS: Introduction to relational databases and MySQL - SQL

Introduction to MySQL ● Introduction ● Installation ● SQL ● Schema design ● Perl BITS/VIB Bioinformatics Training – Jun 24, 2011 – Luc Ducazu <luc@daphnia.com>

SQL ● SQL (Structured Query Language) is the language a RDBMS provides for: – Data definition (DDL) CREATE TABLE, DROP DATABASE – Data manipulation (DML) SELECT, INSERT, DELETE, UPDATE – Data control (DCL) GRANT, REVOKE ● SQL is a ANSI/ISO standard MySQL implements a broad subset of ANSI SQL 99

The MySQL monitor (again) ● There are a number of ways to execute SQL statements using the MySQL monitor: – Interactively $ mysql [database] mysql> stmt; – From the command line $ mysql [database] e 'stmt' – From a file or a pipe (stdin) $ mysql [database] < stmt_file $ cat stmt_file | mysql [database]

Creating a Database ● Only database users with sufficient privileges can create databases (eg root@localhost) ● From the command line: $ mysqladmin [opt] create dbname mysqladmin shares the same command line options for user, password, host and port as mysql ● From within the MySQL monitor: mysql> create database dbname

Exercises ● As root@localhost, create database 'biodb' ● Grant all privileges to the database user you created before: mysql> grant all on biodb.* to user@localhost; ● Download the file biodb1.sql you might want to take a look at the contents (!) ● Execute all SQL statements in this file

Hierarchy ● A single MySQL service can have multiple databases mysql> show databases; ● A particular database db can have multiple tables mysql> use db; mysql> show tables; ● A particular table tbl can have multiple columns or fields mysql> show columns from tbl; mysql> show create table tbl;

Exercises ● Connect to the database service as a normal user ● What databases do you see? ● What tables are defined in biodb? ● What are the column names?

Retrieving rows ● To read data from the database, you use the select SQL statement: select columnlist from src; ● With – columnlist: ● list of columns separated with a comma ● use * to retrieve all columns – src: ● single table or multiple tables joined together ● subquery ● view

Retrieving rows examples ● To show all data from table modorg: select * from modorg; ● To show all model organisms: select genus, species from modorg;

Sorting rows ● When using select statements, the data is displayed in no particular order ● To sort on one or more columns, use the order by col1 [asc|desc] [, col2 [asc|desc]...] clause ● With – colX: a column or column alias – asc: ascending (default) order – desc: descending order

Execcises ● Show the names (genus & species) of all model organisms in the order of the publishing date of the draft ● Show the names (genus & species) of all model organisms sorted by the number of chromosomes (most chromosomes on top) and than alphabetically by name

Calculated rows ● You can add columns in a query that calculate some value using other columns of the same row ● A multitude of functions and operators are available: see help in the MySQL monitor ● Examples: mysql> select 6 * 7; mysql> select concat(class, " ", genus) from modorg; mysql> select now();

Calculated rows numbers ● See help numeric functions ● Operators: +, , *, /, % ● Functions: – sqrt(x), power(x, y), ... – exp(x), ln(x), ... – sin(x), cos(x), ... – round(x), ceil(x), floor(x), ... – rand(), rand(x)

Calculated rows strings ● See help string functions ● Functions: – length(s) – concat(s1, ...) – upper(s), lower(s) – trim(s), ltrim(s), rtrim(s) – substr(s, ...) – reverse(s)

Calculated rows dates ● See help date and time functions ● Functions: – currentdate(), now() – year(d), month(d), week(d) – dayofmonth(d), dayofweek(d) – hour(d), minute(d), second(d)

Exercises ● Show – model organism full name (as a single column) – average chromosome size – publication year of all rows sorted by average chromosome size (largest size on top)

Column aliases ● Columns can be renamed (aliased): select col [as] alias ... ● The aliases can be used in the order by clause

Exercises ● Show – model organism full name (as a single column) as name – average chromosome size as avgsize – publication year as pubyear of all rows sorted by avgsize (largest size on top)

Filtering rows ● To select only those rows that meet certain criteria, use the where clause: select columnlist from src where condition(s) [order by sortcol]; ● With condition(s): – one or more conditions, combined with not, and, or, xor – only the rows for which the condition(s) evaluate(s) TRUE are selected – you can not use column aliases in condition(s)

Filtering rows conditions ● See help comparison operators ● Numerical comparison operations: – = – != or <> – <, <=, >, >= – between x and y (inclusive) ● Example: select all organisms with more than 10 chromosomes: select genus, species from modorg where nchr > 10;

Filtering rows conditions ● String comparison operations: – = – != or <> – <, <=, >, >= (lexical) – like “pattern” matches a pattern: ● _ (a single character cfr shell ?) ● % (0 or more characters cfr shell *) – rlike “regex” [MySQL] matches a regular expression ● Example select all mammals: select genus, species from modorg where class = “mammals”;

Filtering rows conditions ● Dealing with NULLvalues – Testing for NULLness: select ... where col is null ... select ... where col is not null ... – Substitution of NULLvalues: select ifnull(col, value) ... this function returns: ● col if col is not NULL ● value if col is NULL Example: select genus, species, ifnull(nchr, 0) from modorg;

Filtering rows conditions ● Boolean logic: – not x evaluates TRUE if x is FALSE – x and y evaluates TRUE if both x and y are TRUE – x or y evaluates TRUE if x or y is TRUE, or both – x xor y (eXclusive OR) evaluates TRUE if either x or y is TRUE, but not both

Exercises ● Select all mammals with genomes published after 2005 ● Select all organisms that have an average chromosome size between 10 and 100 Mbp ● Select all organisms whose genus starts with A, B, C, D, or E

Filtering rows duplicates ● To eliminate duplicate rows, use select distinct cols from ... ● Each combination of cols is unique

Filtering rows limiting output ● To limit the number of rows in a result set, use select ... limit n [offset r] ● The result set is limited to a maximum of n rows ● If an offset r is given, the first r rows are skipped ● It usually makes little sense to use limit without order by

Exercises ● Give an overview of all organism classes in the dataset (sorted alphabetically) ● Show the organism names of the top 3 largest genome sizes

Aggregation ● So far, queries have concentrated on particular rows ● Sometimes, you need to calculate a single result across multiple rows eg what is the maximum genome size ● SQL allows you to – specify criteria to group rows together – calculate a single value per group – filter grouped data

Aggregation functions ● See: help functions and modifiers for use with GROUP BY ● Functions: – count(col), count(*), count(distinct col) – sum(col) – min(col), max(col) – avg(col), stddev(col), variance(col)

Exercises ● All queries below return a row count. What is the result? Why? – select count(*) from modorg; – select count(nchr) from modorg; – select count(class) from modorg; – select count(distinct class) from modorg; ● How many mammals are in the database?

Aggregation groups ● The group by clause sorts data into groups for the purpose of aggregation: select [col,] aggregatefunctions from src [where cond] group by col [order by ...]; ● All rows with the same value in col are grouped ● For each group, the aggregate function is calculated ● It usually makes no sense to select any columns other than col

Exercises ● How many organisms are present in the dataset for each class? Note the sort order. ● Show the minimum and maximum genome sizes for each class. Take only those organisms into account for which the genome sizes are known. Sort the results such that the biggest maximum genome size is on top.

Aggregation filtering ● It is possible to query filter results based on the results of aggregate functions, using the having clause: select [col,] aggregatefunctions from src [where cond1] group by col having cond2 [order by ...] ● Column aliases can be used in cond2

Execution order 1> Input columns are determined 2> where: input columns are filtered 3> group by: sorting & grouping of filtered input 4> aggregation functions are calculated 5> having: aggregation results are filtered 6> order by: output is sorted 7> limit / offset: output is chopped

Exercises ● For each class with more than 1 organism, show the average number of chromosomes. Sort the result such that the biggest average is on top.

Database upgrade ● To prepare for the next part: – Download biodb2.sql – Delete the database biodb: mysql> drop database biodb; – Recreate the database biodb: mysql> create database biodb; – Create the tables and insert the data $ mysql biodb < biodb2.sql

Joins ● Relational databases model entities and their relationships ● Different entities: different tables ● Joins allow you to combine information across different tables

Joins ● modorg.class_id is a foreign key that references class.id ● gene.mo_id is foreign key that references modorg.id

Joins the Cartesian product ● When you specify multiple tables in a query, the database server will generate all possible combinations: the Cartesian product ● Example: – table class has 6 rows – table modorg has 10 rows – query select * from modorg, class has 60 rows

Joins ● One way to look at joins is to consider it a filtered (think: where) Cartesian product: select * from modorg, class where modorg.class_id = class.id ● Note: – column name id is present in both tables class and modorg to avoid ambiguity, the column name is qualified: class.id – column name modorg.class_id is also qualified, but since there is no ambiguity, this is not strictly necessary

Joins ● Alternative syntax: select * from modorg [inner] join class on modorg.class_id = class.id; ● Both syntactical forms are considered completely equivalent there is no compelling reason to prefer one above the other, it is just a matter of taste.

Joins column aliases ● When specifying columns in the query, care must be taken to avoid ambiguity: mysql> select id, name, genus, species from modorg, class where modorg.class_id = class.id; ERROR 1052 (23000): Column 'id' in field list is ambiguous ● Ambiguous columns must be qualified. And additionally you can choose an alias: mysql> select modorg.id as mo_id, name, genus, species from modorg, class where modorg.class_id = class.id;

Joins data source aliases ● In some cases, you need to join a table with itself or you select data from a subquery ● In this case it can be handy (or even necessary) to alias the data source: select a.col, b.col from src1 [as] a, src2 [as] b where ...

Joins data source aliases ● Consider the (at first sight) simple question: For each class, give the class name, organism name and date of the organism that was sequenced first ● A part of the answer is this: select class_id, min(draft) as dr from modorg group by class_id; ● To add the class name: join with table class ● To add the organism name: join with table modorg

Joins data source aliases ● Add the class name: select name, dr from (select class_id, min(draft) as dr from modorg group by class_id) as s, class where s.class_id = class.id; ● Add the organism name: select name, genus, species, dr from (select class_id, min(draft) as dr from modorg group by class_id) as s, class, modorg where s.class_id = class.id and s.dr = draft;

Exercises ● For all rows in table gene, show – organism name – class name – accession – length – description of the gene

Reusing queries views ● This last exercise is actually very useful as a data source itself: – It could be used as such (as a subquery) but this would end up being inconvenient. – The query as such can be saved as a special tablelike object, called a view. ● Syntax: create view viewname as select ...

Reusing queries views ● You can refer to a view by specifying its name in the from clause: select ... from viewname where ... order by ... ● Although there are circumstances where you can change a view, most views are to be considered as readonly data sources

Exercises ● Create a view (genevw) from the query in the previous exercise ● Select all genes containing hemoglobin in the description and sort the result set by gene length. Next: – What is the minimum and maximum gene length? – What is the average gene length? – And the standard deviation? ● Does the view show up when using mysql> show tables;

Database backup ● You can dump a complete database into a text file: $ mysqldump [opt] db > db.sql ● This includes: – statements for creating the database if the option databases is used – statements for creating tables, views, ... – statements for inserting data ● To restore the database (you may need to create it first): $ mysql db < db.sql

BITS: Introduction to relational databases and MySQL - SQL

More Related Content

What's hot

Viewers also liked

Similar to BITS: Introduction to relational databases and MySQL - SQL

More from BITS

Recently uploaded

BITS: Introduction to relational databases and MySQL - SQL