Workshop offered by Guillaume Larocque (guillaume.larocque@mcgill.ca), research professional at the Quebec Centre for Biodiversity Science.

Link to Prezi presentation

PDF version of the Prezi presentation

• PostgreSQL Reference Manual
• Beginning databases with PostgreSQL
• PostgreSQL wiki
• PGSearch to search PostgreSQL related websites and lists.

• Visit the EntrepriseDB site and download the version appropriate for your platform. Opt to download the 64bit version.
• Execute the file and go through the installation steps, accepting default options.
• When prompted, choose an appropriate password for the postgres user and write this password down.

In a terminal, type:

sudo apt-get install postgresql-9.5 libreoffice libreoffice-sdbc-postgresql

It is possible that the version of postgresql available in your repositories is earlier than 9.5. You can install the version you have available.

For other distributions, check this page.

Installing MySQL

• Click here, download and install LibreOffice 5 (Fresh).

On Windows and Mac:

• Find PostgreSQL 10.X in the Start menu and click on “SQL shell (psql)”
• In the shell, accept the default values by pressing enter for the username and server name and type the proper Postgres password (chosen above). Note that you will not see anything while typing the password. This is normal.

If that doesn't work...

sudo launchctl unload /Library/LaunchDaemons/com.edb.launchd.postgresql-9.5.plist

sudo launchctl load /Library/LaunchDaemons/com.edb.launchd.postgresql-10.plist

On Linux Simply type:

psql

in a terminal.

If that doesn't work...

createuser -P -s -e your_linux_username

psql

SHOW hba_file;

local all postgres peer

local all postgres md5

psql worshop -U your_username -W

If that doesn't work on Mac...

sudo launchctl unload /Library/LaunchDaemons/com.edb.launchd.postgresql-9.4.plist

sudo launchctl load /Library/LaunchDaemons/com.edb.launchd.postgresql-9.4.plist

Click here to download the original spreadsheet

If desired, you can create a superuser other than the default postgres user. To do so, type the following command, replacing your_username and your_password with values of your choice.

CREATE USER your_username WITH SUPERUSER PASSWORD 'your_password';

Creating a new database for the workshop

CREATE DATABASE workshop;

Important step to secure access to database.

REVOKE CONNECT ON DATABASE your_database FROM PUBLIC;

CREATE USER user_name WITH PASSWORD 'your_password';
GRANT ALL ON DATABASE your_database TO user_name;

• Open Libre Office Base and “New…database”.
• Connect to an existing database>PostgreSQL
• For Datasource URL, type the following:

dbname=workshop host=localhost port=5432

• Click Next and put the appropriate user name, and click on 'password required'. You can then click on 'Test connection' to make sure the connection works properly.
• You can then same the database connection information as a file on your computer for easily connecting to it later.

List all databases

\l

Connect to a database

\c workshop

List and describe all tables in the current database

\dt

… with more details

\dt+

View a list of all tables in database

\d+

Describe a particular table

\d tablename

List all users

\du

Quit Postgresql

\q

One you have connected to your database using the instructions above, you can click on Tool…SQL in LibreOffice Base and type your command there. If you create or delete a new table, you might have to click on View…Refresh Tables to see the changes.

More common data types

Full list of PostgreSQL data types.

CREATE TABLE bird_sightings (sight_id SERIAL PRIMARY KEY,
obs_time timestamp,species text,sp_count integer,lat real,"long" real);

We use “” when we want to store a column or table name that is a reserved word or that contains spaces or special characters.

MySQL

CREATE TABLE bird_sightings (sight_id int(5) NOT NULL AUTO_INCREMENT PRIMARY KEY,
obs_time datetime,species text,Sp_count int(10),Lat decimal(8,5),`Long` decimal(8,5));

To add a new row with values to the table

INSERT INTO bird_sightings (obs_time,species,sp_count,lat,"long") VALUES ('2012-06-19 12:31:16','Sturnus vulgaris',40,45.3522,-73.7930);

Note that we did not specify the sight_id field since it is set to SERIAL (auto-increment) and it will be assigned a value automatically.

To see the table:

SELECT * FROM bird_sightings;

If you need to delete the table

DROP TABLE bird_sightings;

Question 1 Reproduce the following table in PostgreSQL, name it 'actors' and visualize it in PostgreSQL

Download the following files to your computers (right-click, save link as…): Lakes.csv lakes_species.csv species_acro.csv

On Windows: Copy those three files to a folder named 'opendbwork' in the C: folder (i.e. C:\opendbwork) On Mac or Linux: Copy those three files to a folder in your path that you can easily identify (e.g. /home/myname/opendbwork). Modify the commands below accordingly by replacing "C:/opendbwork" with your path.

Create the table containing the environmental information of each lake - Lakes

CREATE TABLE lakes (lake_id SERIAL PRIMARY KEY,numero INT, lake_name text, 
province text, latitude DEC(10,5),longitude DEC(10,5), number_of_species INT, ecoprov VARCHAR(50), 
ecozone VARCHAR(50), gss DEC(10,2),gse DEC(10,2), gsl DEC(10,2), gdd10 DEC(10,2), egdd DEC(10,2), 
mean_ele DEC(10,2), pe_ann_p DEC(10,2), totp_ann DEC(10,2), srann_me DEC(10,2), shann_me DEC(10,2), 
tmax_ann DEC(10,2), tmean_an DEC(10,2), vpann_me DEC(10,2));

MySQL

CREATE TABLE lakes (lake_id INT NOT NULL AUTO_INCREMENT PRIMARY KEY,numero INT, lake_name text, 
province text, latitude DEC(10,5),longitude DEC(10,5), number_of_species INT, ecoprov VARCHAR(50), 
ecozone VARCHAR(50), gss DEC(10,2),gse DEC(10,2), gsl DEC(10,2), gdd10 DEC(10,2), egdd DEC(10,2), 
mean_ele DEC(10,2), pe_ann_p DEC(10,2), totp_ann DEC(10,2), srann_me DEC(10,2), shann_me DEC(10,2), 
tmax_ann DEC(10,2), tmean_an DEC(10,2), vpann_me DEC(10,2));

Load the CSV file into this table

COPY lakes FROM 'C:/opendbwork/Lakes.csv' WITH csv HEADER DELIMITER AS ',';

If this command gives you an access denied error message in Windows

If this command gives you an access denied error message on Mac

MySQL

LOAD DATA LOCAL INFILE 'C:/opendbwork/Lakes.csv' INTO TABLE lakes FIELDS TERMINATED BY ',' 
ENCLOSED BY '"' LINES TERMINATED BY '\n'   IGNORE 1 LINES;

mysql -u root -p --local-infile=1;

Create the table containing the Species presence at each lakes and load CSV into it - lakes_species

CREATE TABLE lakes_species (lake_id INT NOT NULL, species_id INT NOT NULL);
COPY lakes_species FROM 'C:/bdlibre/lakes_species.csv' WITH csv HEADER DELIMITER AS ',';

MySQL

CREATE TABLE lakes_species (lake_id INT NOT NULL, species_id varchar(25) NOT NULL);
LOAD DATA LOCAL INFILE 'C:/opendbwork/Lakes_Species.csv' INTO TABLE lakes_species 
FIELDS TERMINATED BY ',' ENCLOSED BY '"' LINES TERMINATED BY '\n'   IGNORE 1 LINES;

For the third table, we could use the command line with the following:

Command

CREATE TABLE lakes_species (lake_id INT NOT NULL, species_id INT NOT NULL);
COPY lakes_species FROM 'C:/bdlibre/lakes_species.csv' WITH csv HEADER DELIMITER AS ',';

However, we will use LibreOffice instead…

• Open the Species_Acro.csv file with LibreOffice Calc (In Windows, find LibreOffice Calc in the start menu). Specify that the file is separated by commas and specify utf8 as the character set (encoding).
• Select the entire worksheet with ctrl-a;
• Open LibreOffice Base and go to Edit…paste. Call the new table public.species_acro. Select 'Definition and data'. Select to use first line as column names. Do not create a primary key at this point.
• Follow the wizard… Move all columns to the right to select them. Then, specify the proper data format (verify the types of attributes in the table) for the other columns. Right click on the column named 'species_id' and select 'Primary Key'. Set the unique_id as int [int4], full_name as text [varchar] with 200 characters and species_id as text [varchar] with 30 characters.

MATH AND GROUP BY (Aggregate) Functions

Full list

Conditional selections used in the where clause

Order of operators in a SELECT statement:

SELECT columns FROM tables WHERE conditions JOIN GROUP BY columns HAVING condition ORDER BY columns;

Select all rows from the lakes table.

SELECT * FROM lakes;

In the terminal, when you are done browsing a table, hit 'q' to go back to the prompt.

Select just the lake names and province.

SELECT lake_name, province FROM lakes;

Show all province names in table

SELECT DISTINCT province from lakes;

Show all province names in table, use an alias for province and order by descending alphabetical order.

SELECT DISTINCT province as "province name" FROM lakes ORDER BY province DESC;

Show the birth year of each actor

SELECT first_name, last_name,  EXTRACT(YEAR FROM birth_date) FROM actors;

Show initials of each actor

SELECT first_name, last_name, concat(substr(first_name,1,1),' ',substr(last_name,1,1)) as initials FROM actors;

Select all lakes from Quebec where the mean annual temperature is below -5.

SELECT lake_name, tmean_an FROM lakes WHERE province='QUEBEC' AND tmean_an<-5;

Or simply, to count the number of lakes

SELECT count(*) FROM lakes WHERE province='QUEBEC' AND tmean_an<-5;
or equivalently:
SELECT count(lake_id) FROM lakes WHERE province='QUEBEC' AND tmean_an<-5;

Question 1 - How many lakes in British Columbia receive more than 3000 mm of precipitation (totp_ann column)?
answer 12
Question 2 - What is the average Elevation (mean_ele column) of all lakes in the Montane Cordillera Ecozone (use the avg() operator)?
answer 1364.36

Note: the % operator is a wildcard used to replace the beginning or the end of a string.
Select all lake names that contain the word 'Small'

SELECT lake_name FROM lakes WHERE lake_name like '%Small%';

Select all lake names that contain the word 'Small' and that are not located in Ontario.

SELECT lake_name, province 
FROM lakes
WHERE lake_name like '%Small%' AND province!='ONTARIO';
other possibility:
SELECT lake_name, province 
FROM lakes 
WHERE lake_name like '%Small%' AND province NOT LIKE 'ONTARIO';

Question 3 - What is the maximum latitude of all lakes in Ecozones with names that start with 'Taiga'
answer 68.8
Question 4 - How many species of Daphnia are there in the species_acro table?
answer 17 (21 have the word 'Daphnia' somewhere in their name)

Calculate the average annual temperature of all lakes within each province

SELECT province, AVG(tmean_an) as mean_an_t
FROM  lakes
GROUP BY province;

Same table ordered by increasing temperatures

SELECT province, AVG(tmean_an) as Mean_AN_T
FROM  lakes
GROUP BY province ORDER BY AVG(tmean_an);

Question 5 - Which province has the lake with the highest maximum temperature (column tmax_ann)?
answer BRITISH COLUMBIA 14.60

Update statement: Give all rows where the province name includes NWT the same province name 'NWT' - Careful, this actually modifies the table!

UPDATE lakes
SET province='NWT'
WHERE province like '%NWT%';

Count the number of species in each lake. Display lake name and species number.

SELECT lake_name, COUNT(DISTINCT species_id) as num_species FROM lakes, lakes_species 
WHERE lakes.lake_id=lakes_species.lake_id 
GROUP BY lakes.lake_id 
ORDER BY num_species DESC;

Same query, but only display the first 20.

SELECT lake_name, COUNT(DISTINCT species_id) as num_species FROM lakes, lakes_species 
WHERE lakes.lake_id=lakes_species.lake_id 
GROUP BY lakes.lake_id 
ORDER BY num_species DESC LIMIT 20;

Count the number of species in each Ecozone and Ecoprovince. Display lake name and species number.

SELECT ecozone, ecoprov, COUNT(species_id) as num_species FROM lakes, lakes_species 
WHERE lakes.lake_id=lakes_species.lake_id 
GROUP BY ecozone, ecoprov
ORDER BY ecozone, COUNT(species_id);

Question 6 Generate a table that lists the full name of each species of Daphnia and the number of lakes where this species is present?

Answer

SELECT full_name, count(DISTINCT a.lake_id) as sp_count FROM lakes a, species_acro b, lakes_species c WHERE full_name like 'Daphnia%' AND a.lake_id=c.lake_id AND b.species_id=c.species_id GROUP BY full_name;

Create a list of all the species found in the lakes in Quebec, with the full names of the species and the number of occurences

SELECT full_name, count(full_name) 
FROM species_acro,lakes_species,lakes 
WHERE species_acro.species_id=lakes_species.species_id AND lakes.lake_id=lakes_species.lake_id AND province='QUEBEC' 
GROUP BY full_name;

Alternatively, and equivalently…

SELECT full_name, count(full_name) 
FROM lakes
LEFT JOIN lakes_species ON (lakes.lake_id=lakes_species.lake_id)
LEFT JOIN species_acro ON (species_acro.species_id=lakes_species.species_id)
WHERE province='QUEBEC' 
GROUP BY full_name;

For every possible combinations of lakes found in the Lake of the Woods Ecoprovince, return the difference in mean annual temperature between those lakes.

SELECT concat(a.lake_name,'-',b.lake_name) as lakes, a.TMEAN_AN-b.TMEAN_AN as temp_diff 
FROM lakes a, lakes b 
WHERE a.ecoprov='Lake of the Woods' AND b.ecoprov='Lake of the Woods';

Generate a table containing the count of the presence of each species in lakes North and South of the 50th parallel (latitude).

SELECT a.species_id, count_50south, count_50north
FROM 
 (SELECT species_id, count(d.species_id) as count_50south FROM lakes c, lakes_species d WHERE c.lake_id=d.lake_id AND latitude<=50 GROUP BY species_id) a,
 (SELECT species_id, count(f.species_id) as count_50north FROM lakes e, lakes_species f WHERE e.lake_id=f.lake_id AND latitude>50 GROUP BY species_id) b 
WHERE a.species_id=b.species_id;

Question 7 Using a subquery, find the mean elevation of lakes (MEAN_ELE column) where are species of Daphnia is present
Answer 474.627172
Code: SELECT avg(mean_ele) FROM (SELECT DISTINCT a.lake_id, mean_ele FROM lakes a,species_acro b,lakes_species c WHERE b.full_name like 'Daphnia%' AND a.lake_id=c.lake_id AND c.species_id=b.species_id ORDER BY a.lake_id) d;

Create a new table with just lakes in Quebec

CREATE TABLE lakes_qc AS SELECT * FROM lakes WHERE province='QUEBEC';

COPY lakes_qc TO 'C:/opendbwork/lakes_qc.csv' WITH csv HEADER DELIMITER AS ',';

• Click on Forms>Create Form in Design view.
• Click on the icon associated with the first form element you want to add to the form. For example, click on Text Box and then click/drag to delineate the location of the element.
• Right-click on the form element and click on Form… and then click on the Data tab. In the Content drop-down menu, choose the main table that you want to fill with this form. You only need to select this once for the whole form. Close this dialog.
• Right-click again on the form element and click on Control… Data and choose the column that you want to fill through this form element.
• You can keep adding other elements in the same way, repeating the previous step each time.
• To test your form, save it, close the Design view and open your form by double-clicking on its name.

• For a dropdown/select menu that contains elements from a secondary table, click on the List Box icon and place it on the form. Then, a wizard will appear to ask you which table and column should be used to show options of the menu. Then, you will have to select the foreign and primary keys from the two tables. When an option is selected through this menu, the foreign key of this option will me added to the associated column in the main table.

• To have radio buttons that are mutually exclusive, add them to your form and then right-click on each of them and click on Control… Associated each of them with the same Data field, but specify a different Reference value(on) for each. Leave the (off) value blank.
• Then, select the radio buttons and right-click and select Group.

You are given a dataset in the following format:

You are told that thousands more trees will have to be entered in this way. How would you go about designing a database and form system to store this information?

Create the following table in PostgreSQL. You can then fill the table with its values by clicking on the table name in Base and entering values manually. Table name: tree_species

Create this table structure and create a form to fill it in the form design view, with a drop-down menus (list box) to select the species names and radio buttons for the tree health. tree_id is the primary key. For technical reasons, make sure that you specify varchar as the format for the status field. Table name: tree_plots

• On Windows, download the PostgreSQLODBC driver and install it on your computer.
• Create your tables in PostgreSQL.
• In Access, click on External data… ODBC database, and then select “Link to the data source by creating a linked table” and then choose New DSN name. Specify a name for the file, and then fill the information to connect to the database. You can then choose the tables you want to link.
• Those table will then be linked to PostgreSQL and you can edit then as you would with other Access tables.

• Access the Prezi presentation here

Useful link

• Code Academy Great website to learn HTML, CSS, PHP, Javascript, JQuery or other.
• Beginning PHP and PostgreSQL 8. Book that can be downloaded in e-book format. A bit outdated.

Main steps

1. Creation of a web interface with html code and CSS.
2. Creation of a database to accept the form data.
3. Creation of PHP code to receive the form data and send it to the database.
4. Possibility to enrich the form with Javascript/JQuery

Download and install the WAPP server for pour Windows, or MAPP for Mac.

For linux:

sudo apt-get install postgresql-9.5 apache2 php5 phppgadmin php5-pgsql

Main components needed for an HTML file

<!DOCTYPE html>
<html>
<head>
<title>Document title</title>
</head>
<body>
The page content is here!
</body>
</html>

Creation of a form

The basic form

<form method="POST" action="phpfile.php">
<input type="text">
</form>

List of html input types

Text input

<input type="text" name="name_text">

Dropdown menu

<select name="select1">
<option value="1">Option 1</option>
<option value="2">Option 2</option>
<option value="3">Option 3</option>
</select>

Check box

<input type="checkbox" name="poutine" value="yes"> I like poutine
<input type="checkbox" name="sushis" value="yes"> I like sushi
<input type="checkbox" name="poutine_sushis" value="yes"> I like poutine topped with sushi

Radio buttons

I am a...
<input type="radio" name="gender" value="m" > Man
<input type="radio" name="gender" value="w"> Woman

Text area

<textarea name="text_zone">
</textarea>

CSS reference

You can give an “id” to any html element, but there can only one element per id.

<input type="text" name="fieldname" id="fieldid">

We can then reference it in the header of the css file with the pound symbol (#)

<head>
<style>
#fieldid {
 width:300px;
 height:30px;
 color:red;
 background-color:#ffff00;
 font-family:verdana,arial;
 font-size:14px;
 font-weight:bold;
 padding:0px 5px 0px 5px; #top right bottom left
}
</style>
</head>

We can also give a class to any html element. This class can be repeated for multiple elements.

<input type="text" name="name_text1" class="class_text">
<input type="text" name="name_text2" class="class_text">
<input type="text" name="name_text3" class="class_text">

Et on peut y faire référence dans l'entête de style css du fichier html avec le point “.” We can then reference it in the style section of the html header with the “.”

<head>
<style>
.class_text {
 width:300px;
 height:30px;
 color:red;
 background-color:#ffff00;
 font-family:verdana,arial;
 font-size:14px;
 font-weight:bold;
 padding:0px 5px 0px 5px; #top right bottom left
}
</style>
</head>

Link to phppgadmin

All php commands must be contained within

<?php
PHP code is here
?>

Comments

<?php
// Comments in a line
/*
Comments section
*/
?>

<?php
// Creation of variables
$myvar=5;
$myvar='this is text';
 
// Creation of an array
$myarray=array('value1','value2','value3');
 
echo $myarray[0];
//value1
 
$myarray=[];
$myarray[0]='value1';
$myarray[1]='value2';
 
// Creation of a "key-value" array
$myarray=array(
'key1'=>'key1',
'key2'=>'key2',
'key3'=>'key3',
);
 
echo $myarray['key1'];
//value1
?>

Loops and if/else

<?php
if ($f==5){
//Do something
}else{
//Do something else
}
 
for ($i=0;$i<10;$i++){
echo 'i='.$i;
}
//i=0
//i=1
//...
//i=9
 
 
foreach ($myarray as $my){
//Do something here with $my
?>

PostgreSQL functions in PHP

This PHP file is the one which is specified in the “ACTION” part of the <form> field

<?php
// Connexion and selection of the database
$dbconn = pg_connect("host=localhost dbname=database user=username password=my_password");
 
// The $_POST contains all the data from the form. Each 'key' in this array corresponds to the name given to each html element. 
$element1=$_POST['name_element1'];
$element2=$_POST['name_element2'];
$element1=pg_escape_string($element1);
$element2=pg_escape_string($element2);
 
// Inserted into the databse
$query = "INSERT INTO ma_table (column1,column2) VALUES ($element1,$element2)";
$result = pg_query($query);
// Close the connexion
pg_close($dbconn);
echo "Form successfully entered into the database|";
?>

Accéder à différents types d'éléments

//Check boxes
foreach ($_POST['checkbox'] as $check){
   //Faire quelque chose ici avec $check
}
//Convertir les valeurs en liste séparée par des virgules
$check=implode(',',$_POST['checkbox']);

Retrouver des éléments dans la base de données

// Connexion et sélection de la base de données
$dbconn = pg_connect("host=localhost dbname=base_donnees user=nom_usager password=mot_de_passe");
 
$query="SELECT * FROM nom_table";
// Insertion dans la base de données
$result = pg_query($query);
while ($t=pg_fetch_array($result)){
	echo $t['prenom'].' '.$t['nom_famille'];
}
 
// Fermer la connexion
pg_close($dbconn);

From within R: On first use only:

install RPostgreSQL package

install.packages('RPostgreSQL')

R Code

library(RPostgreSQL)
drv <- dbDriver("PostgreSQL")
con <- dbConnect(drv, user="your_username", password="your_password", dbname="workshop");
lakes <- dbGetQuery(con,'SELECT * FROM lakes');

Plot some data

lakesqc <- dbGetQuery(con,"SELECT * FROM lakes WHERE province='QUEBEC'")
hist(lakes$tmean_an)

library(dplyr)
src<-src_postgres(dbname="workshop",host="localhost", port="5432",user="your_username",password="your_password")
lakes <- tbl(src, "lakes") # Define lakes table
lakes_qc<-filter(lakes, province  %=% 'QUEBEC') # Select lakes in Quebec
prov_tmean<-summarise(group_by(lakes, province), mean(tmean_an)) # Mean annual temperature per province
prov_tmean=collect(prov_tmean) # Transfer result to standard R data frame
lakes_qc2<-tbl(src, sql("SELECT * FROM lakes WHERE province='QUEBEC'")) #Perform any SQL statement