SQL CHEAT-SHEET PAGE 1

LEFT JOIN
1 ATA B ASE
D
dml commands
BETWEEN The BETWEEN operator can be used to filter SELECT * FROM table_name LIMIT 5 ;
by a range of values.
G RANT SELECT , UP DATE ON TABLE_1
--Limit is used to limit the result set SELECT col1, col2

CREATE DATABASE database_name; TO USER_1, USER_2; FROM table_name t1

-- reates a new database LIKE The LIKE operator can be used inside of a to the specified number of rows.
C
--Used to grant a user access WHERE clause to match a specified pattern.
LEFT JOIN table_name t2

U SE database_name; privileges to a database.

The % wildcard can be used in a LIKE
SELECT col1, col2 FROM table_name
ON condition;
--Uses the specified database
%

Wildcard operator pattern to match zero or more ORDER BY col1 ASC [ DESC ] ; TABLE 1 TABLE 2
G RANT SELECT , UP DATE ON TABLE_1
unspecified character(s). --Sort the result set in ascending or L
TO USER_1, USER_2; The _ wildcard can be used in a LIKE
descending order 1 E
F
A 1 B
CO MM AN S --Used to grant a user access
2 sql
=
_

privileges to a database. Wildcard operator pattern to match any single

unspecified character. SELECT col1, col2 FROM table_name
2 T

J
B 2 A
database ORDER BY col1 LIMIT n OFFSET offset; O
CREATE TABLE table_name
dcl commands
IN
The IN operator is used to compare the
specified value. AS - Columns or tables can
--Skip offset of rows and return the 3 I
N C 3
( column1 datatype,
c o mmi t Saves all the transactions made on a be aliased using the AS clause. next n rows based on LIMIT.
Returns all rows from the table on the left side of the
column2 datatype,
database.
It is used to search for the presence of a row SELECT col1, aggregate (col2)
JOIN and matched rows from the right side of the Join
column3 datatype); E IST
It is used to undo transactions which are
FROM table_name G RO UP BY col1;
X
ro llb a ck in a table.
--The create table statement creates not yet been saved.
--GROUP BY Groups rows using an aggregate R I GH T JOIN
a new table in a database tcl commands
where
function
SELECT column_name FROM table_name
SELECT col1, col2

ALTER TABLE table_name

SELECT col1,col2.. FROM table_name; WHERE column_name IS N U LL ;
SELECT col1, aggregate(col2)
FROM table_name t1

ADD column_name datatype; --Retrieve data from specified columns -- olumn values can be or have no FROM table_name
RI G HT JOIN table_name t2

in the table G RO UP BY col1 HA V IN G condition;

C NULL

--The Alter table statement is used value. These records can be atc ed using m h ON condition;
to odify the columns of an existing
m
SELECT * FROM table_name; the and operators. --Filter groups using the HAVING clause. TABLE 1 TABLE 2
table and add a new column.
IS NULL IS NOT NULL
R
--Retrieve the data from all fields in union 1 I
G A 1 B
ALTER TABLE table_name
the table. SELECT col1, col2 FROM table_name
DESC table_name; H
ADRO P COL U MN column_name; SELECT col1,col2..FROM table_name
U NION --Describes the structure of the table. 2 T

J
O
B = 2 A
--The Alter table statement is used WHERE condition; SELECT col1, col2 FROM table_name;
3 I
N C C
to odify the columns of an existing
m
--Used to filter the records based --Co bine rows from two queries wit out 10 CASE
table and Drop column. on a particular condition.
m

any duplicates.
h

Returns all rows from the table on the right side of the
ALTER TABLE table_name
SELECT column_name,
JOIN and matched rows from the left side of the Join
RENAME col_name TO col_newname; 7 S Q L CONST R AINTS union all CASE

SELECT col1, col2 FROM table_name WHEN Condition THEN ‘output’

C R OSS JOIN
--Rena es the column names in the
m NOT N U LL Specifies that this column cannot store a
NULL value. U NION ALL
WHEN Condition THEN ‘output’

existing table. SELECT col1, col2 FROM table_name; .

SELECT col1, col2

U NI QU E Specifies that this column can have only

-- ombine rows from two queries with .
FROM table_name t1

D rop ta bl e table_name; unique values C

duplicates .
CROSS JOIN table_name t2

--Drop deletes both structure and P RIMAR Y

ON condition;
records in the table. K EY
It is a field using which it is possible to
uniquely identify each row in a table. intersect
ELSE ‘output’

END ‘new_colname’ FROM table_name; TABLE 1 TABLE 2

T run c ate ta bl e table_name; FOREI GN
It is a field using which it is possible to SELECT col1, col2 FROM table_name
-- runcate deletes the table but not K EY
uniquely identify each row in some other INTERSECT 11 JOINS 1 c
r A 1 A
T table. SELECT col1, col2 FROM table_name;
o
the structure
CHEC K It validates if all values in a column --Return the co on rows of two queries.
mm INNE R JOIN 2
s
s

J
B = 1 B
ddl commands
satisfy some particular condition or not
union SELECT col1, col2
3 O
I C 1 C
DEFA U LT It specifies a default value for a column
FROM table_name t1
N
INSERT INTO table_name
when no value is specified for that field SELECT col1, col2 FROM table_name
INNER JOIN table_name t2

2 A
V AL U ES (value1, value2); MIN U S
8 O P E R ATO R S SELECT col1, col2 FROM table_name;
ON condition; 2 B
UP DATE table_name

SET column1 = value1,column2 = value2

and The NOT operator allows the negotiation of the
--Returns the values from the first table TABLE 1 I TABLE 2 2 C
WHERE some_column =some_value; condition. after re oving t e values from the second 1 N
A 1 B 3 A
m h
N
--The PDA E statement is used to edit or The OR operator allows multiple conditions to be table. E
2 R

B = 2 A 3 B
U T

records (rows) in a table. combined. Records match either condition.

J
Querying
not The AND operator allows multiple conditions to be O
9 D ata

DELETE FROM table_name

combined. Records must match both conditions
SELECT DISTINCT (column_name)
3 N
I C 3 C
WHERE some_column = some_value; ALL It compares a value to all the values in another set. FROM table_name;
--The delete statement is used to -- nique values of the columns are Only returns rows that meet the join condition Produce a Cartesian product of rows in tables
delete records (rows) in a table. ANY It compares the values in the list according to the
condition.
U

retrieved from the table.

 SQL CHEAT-SHEET PAGE 2

FULL JOIN L AST _ anking is assigned to the rows based The implementation of APPROX_COUNT_DISTINCT() Creating a search index:
in d o w

●
Gives the last value to all rows. R

V A LU E ()
w
on a specified column. Ranks are not has a much smaller memory requirement as Indexes are created for one or more columns in
SE L ECT col1, col2
functions skipped. compared to the COUNT(DISTINCT) function a table, using the CREATE INDEX command.

F RO M table_name t 1
last_

Gives the last value to all rows. Syntax: CREATE SEARCH INDEX index_name ON
FULL J OIN table_name t2
value() s e l e ct the database selects the given columns,
lo w er () :
`dataset.table_name` (column_name);
ON condition; istinct
Nth
d repeated values are removed, The LOWER() function converts a string to lowercase.
Gives Nth value to all rows. Dropping a search index:
TABLE 1 TABLE 2 value()
union /
upper () :
B An index can be dropped using SQL DROP command.
1 A 1 NTILE() Divides the rows to ‘n’ number of small int e r s e ct /
the database applies set operations, The UPPER() function converts a string to uppercase.
buckets. e xc e pt Syntax: DROP SEARCH INDEX index_name
2 B = 2 A
The cumulative percentage of the records order the results are sorted,
REGEXP _ C ONT AI N S () :
ON `dataset.table_name`;
cume_
by
Returns TRUE if the value is a partial match for
is calculated from the first row to the ●

3 C C dist() current row for the specified column. O FF SET the first rows are skipped, the regular expression, regex.

18 sql partitioning
3 l imit / ● If the regex argument is invalid, the
Partitioning is a database process where very
14 i
v e ws f e tch / top
only the first rows are selected function returns an error.
large tables are divided into multiple
Returns all rows from bith sides even if join condition is ● You can search for a full match by using ^
smaller, individual parts.

not met SE L ECT V IE W view_name

(beginning of text) and $ (end of text).

AS SE L ECT * F RO M table_name; 16 a d v an c e sql

● By splitting a large table, queries that

12 aggr e gat e fu n c tions --It creates a simple view. Syntax: REGEXP_CONTAINS(value, regex) access only limited amount of data can run
count ( )
faster because there is less data to scan.

avg() returns the average of a list Function with the DISTINCT clause eliminates the REGEXP _ extract () :
SE L ECT V IE W view_name
repetitive appearance of the same data. The DISTINCT

● Returns the first substring in value that

sum() returns the total of a list. AS SELECT col1, col2
can come only once in a given select statement. ● You can partition BigQuery tables by:
matches the regular expression, regex.
count()
F RO M table_name t1
Returns NULL if there is no match.

returns the number of elements of a list. -Time-unit column: Tables are partitioned based on
INNER JOIN table_name t2
count (*)

a TIMESTAMP, DATE, or DATETIME column in the
min() returns the minimum value of a list. ON condition; When * is used as an argument, it simply counts the ● If the regular expression contains a table.

max() returns the maximum value of a list. --It creates a complex view total number of rows including the NULLs.
capturing group, the function returns the -Ingestion time: Tables are partitioned based on
substring that is matched by that capturing the timestamp when BigQuery ingests the data.

count ( 1 )
13 a d v an c e d aggr e gat e fu n c tions CREATE TE M PORAR Y V IE W view_name
group.

With COUNT(1), there is a misconception that it

AS SE L ECT col1, col2
-Integer range: Tables are partitioned based on an
counts records from the first column. What COUNT(1)
t is a window function used inside every F RO M table_name; If the expression does not contain a integer column.

ov e r () I
really does is that it replaces all the records you ●
analytical function --It creates a temporary view. get from query results with the value 1 and then capturing group, the function returns the
Use the CREATE TABLE statement with a SELECT AS
P a r tition counts the rows meaning it even replaces a NULL with entire matching substring.

C reates a partition internally and later clause for the query. Include a PARTITION BY
() 1 meaning it takes NULLs into consideration while

by p erforms the specified operations. DROP V IE W view_name; clause to configure the partitioning.
counting. ● Returns an error if:

r o w _ num b e r
Provides row numbers for all the rows --Delete a view. --The regular expression is invalid
Copying individual partitions:
() based on a specified column in the table. C OUNT ( colu m n _ na m e ) - --The regular expression has more than one ● You can copy the data from one or more

Ranking is assigned to the rows based

15 sql trigg e rs When a column name is used as an argument, it simply capturing group
partitions to another table.

r an k () on a specified column. Skips the rank counts the total number of rows excluding the NULLs
All query elements are processed in a very strict order:
meaning it will not take NULLs into consideration.
Syntax: REGEXP_EXTRACT(value, regex) Copying partitions is not supported by
when it contains the same values. ●
Query execution order. Console.

anking is assigned to the rows based 17 sql in d e x e s

de ns e _
R
COUNT(DISTINCT expr,[expr...])
on a specified column. Ranks are not the database gets the data from tables in
r an k () ● However you can copy them using the API
skipped. f r om FROM clause and if necessary performs
the JOINs,
APPROX _ CO U NT _ DISTINC T
●A SQL index is a quick lookup table used to

Deleting a partition
PERCENT _
Assigns the rank to the specified is useful if an exact result is not required. quickly retrieve data from a database.

r an k () column within the range of 0-1.

w here the data are filtered with conditions ● You can delete an individual partition
specified in the WHERE clause, Compared to SE L ECT from a partitioned table.

● ● Indexes are generally used for large

The first value becomes NULL. CO U NT ( DISTINCT <co l umn> ) , which calculates

L A G () the data are grouped by conditions databases. They are small, fast, and less
C ompares the current value with the g r oup by
the exact number of distinct values in a ● But you can't delete the special NULL or
p revious value. specified in the WHERE clause, memory-consuming.

column of the table, APPROX _ CO U NT _ DISTINCT UN-PARTITIONED partitions.

The last value becomes NULL. Compares A gg r e gat e the aggregate functions are applied to the can run much faster and consume significantly ● Indexing a table or view surely improves
L EAD () functions groups created in the GROUP BY phase,
the current value with the next value. less memory. the performance of queries and applications. ● You can only delete one partition at a

F IRST _ the groups are filtered with the given time

Gives the first value to all rows. H aving condition,
V A LU E ()