AMI Cheat Sheet

AMI Cheat Sheet

For this demonstration we need to include company.db and university.db sqlite databases as datasources into AMI. For better comprehension, the examples should be executed in the order presented

CREATE

CREATE TABLE BusinessTeam (name string, job string, salary int);

CREATE PUBLIC TABLE ManagementTeam (name string);

CREATE TABLE IF NOT EXISTS OpenJobs (job string, id int);

CREATE TEMPORARY TABLE BonusJobs (job string, id int, bonus double); // create temporary table is the same as create table

CREATE TABLE Graduates (name string, degree string);

CREATE AS

CREATE TABLE Team AS USE ds = "company" EXECUTE SELECT * from Employees;

CREATE TABLE SoftwareTeam AS SELECT name, job, salary FROM Team WHERE job IN ("Programming", "Testing", "DevOps");

CREATE TABLE AnalyzeBudget AS ANALYZE name, job, salary, avg(win.salary) AS avgSalary, median(win.salary) AS medianSalary, salary - median(win.salary) AS diff FROM SoftwareTeam WINDOW win ON true;

CREATE TABLE PrepareBudget AS PREPARE name, job, salary, stack(salary), norm(salary) FROM SoftwareTeam ORDER BY salary ASC PARTITION BY job;

CREATE TABLE DescTable AS DESCRIBE TABLE OpenJobs;

CREATE TABLE ShowTable AS SHOW TEMPORARY TABLES;

CREATE TABLE PrefixTeam AS CALL PrefixFind("Jo", "name", "Team");

INSERT INTO

INSERT INTO OpenJobs VALUES ("Programming", 1),("Sales", 2), ("Management", 3), ("Finance", 4), ("DevOps", 5), ("Fax", -1) LIMIT 5;

Double discountFactor1 = 1.03, discountFactor2 = 1.05, discountFactor3 = 1.09, discountFactor6 = 1.08;

Double bonusAmount1 = 25, bonusAmount2 = 20, bonusAmount3 = 30, bonusAmount6 = 15;

INSERT INTO BonusJobs VALUES ("Programming", 1, power(bonusAmount1, discountFactor1)), ("Sales", 2, power(bonusAmount2, discountFactor2)), ("Management", 3, power(bonusAmount3, discountFactor3)), ("Machine Learning,AI", 6, power(bonusAmount6, discountFactor6));

INSERT INTO Graduates VALUES ("Jeff", "Programming"), ("Peter", "Machine Learning"), ("Jennifer", "Finance"), ("Michelle", "Sales");

INSERT INTO ... FROM

INSERT INTO Team FROM SELECT name, degree, 80 FROM Graduates;

INSERT INTO ManagementTeam FROM SELECT name from Team WHERE job == "Management";

INSERT INTO BusinessTeam FROM SELECT name, job, salary FROM Team WHERE job IN ("Management", "Sales", "Operations");

INSERT INTO Team USE ds = "university" EXECUTE SELECT name, subject, 2 * salary FROM Professors;

INSERT INTO AnalyzeBudget FROM ANALYZE name, job, salary, avg(win.salary) AS avgSalary, median(win.salary) AS medianSalary, salary - median(win.salary) AS diff FROM BusinessTeam WINDOW win ON true;

INSERT INTO PrepareBudget FROM PREPARE name, job, salary, stack(salary), norm(salary) FROM BusinessTeam ORDER BY salary ASC PARTITION BY job;

INSERT INTO DescTable FROM DESCRIBE TABLE Team;

INSERT INTO ShowTable FROM SHOW PUBLIC TABLES;

INSERT INTO PrefixTeam FROM CALL PrefixFind("Ma", "name", "Team");

INSERT INTO BYNAME ... FROM

INSERT INTO MiamiHeat BYNAME FROM SELECT * FROM IndianaPacers;

INSERT INTO MiamiHeat BYNAME FROM SELECT *, a AS b FROM IndianaPacers; // manually map column a to column b during the insertion

ALTER

ALTER TABLE Team MODIFY salary AS salary BigInteger;

ALTER TABLE Graduates RENAME degree TO profession;

ALTER TABLE Graduates ADD income int BEFORE profession;

ALTER TABLE Graduates DROP income;

ALTER TABLE Graduates ADD income int;

RENAME

RENAME TABLE ManagementTeam TO Leaders;

DESCRIBE

DESCRIBE TABLE Team;

SELECT

SELECT * FROM Team;

SELECT * FROM Team LIMIT 3;

SELECT * FROM Team LIMIT 3, 7;

SELECT * FROM Team WHERE job == "Programming" OR job == "Testing";

SELECT * FROM Team WHERE job IN ("Finance", "Machine Learning");

SELECT * FROM Team WHERE salary < 100;

SELECT * FROM Team WHERE job == "Programming" AND salary == 100;

SELECT * FROM Team WHERE (job, salary) IN ("Programming", 100);

SELECT * FROM Team ORDER BY salary ASC;

SELECT sum(salary) AS Budget FROM Team;

SELECT job, count(*) AS numberOfEmployees, sum(salary) AS jobBudget, avg(salary) AS jobAvgSalary FROM Team GROUP BY job ORDER BY numberOfEmployees;

SELECT job, count(*) AS numberOfEmployees, sum(salary) AS jobBudget, avg(salary) AS jobAvgSalary FROM Team GROUP BY job HAVING numberOfEmployees >=2;

SELECT * EXCEPT (salary) from Team;

SELECT job FROM Team GROUP BY job;

SELECT * FROM OpenJobs, BonusJobs WHERE OpenJobs.id == BonusJobs.id;

SELECT OpenJobs.job AS job, OpenJobs.id AS id, bonus FROM OpenJobs, BonusJobs WHERE OpenJobs.id == BonusJobs.id;

SELECT * FROM Team WHERE job IN (SELECT job FROM OpenJobs);

SELECT * FROM OpenJobs JOIN BonusJobs on OpenJobs.id == BonusJobs.id;

SELECT * FROM OpenJobs OUTER JOIN BonusJobs on OpenJobs.id == BonusJobs.id;

SELECT * FROM OpenJobs OUTER ONLY JOIN BonusJobs on OpenJobs.id == BonusJobs.id;

SELECT * FROM OpenJobs LEFT JOIN BonusJobs on OpenJobs.id == BonusJobs.id;

SELECT * FROM OpenJobs LEFT ONLY JOIN BonusJobs on OpenJobs.id == BonusJobs.id;

SELECT * FROM OpenJobs RIGHT JOIN BonusJobs on OpenJobs.id == BonusJobs.id;

SELECT * FROM OpenJobs RIGHT ONLY JOIN BonusJobs on OpenJobs.id == BonusJobs.id;

SELECT * FROM BonusJobs UNPACK job ON ",";

UPDATE

UPDATE Graduates SET income = 50;

UPDATE Team SET salary = salary + 5 WHERE job == "Sales";

UPDATE Team JOIN BonusJobs ON BonusJobs.job =~ Team.job SET salary = salary + bonus;

USE ... EXECUTE

USE ds = "company" EXECUTE SELECT * from Employees;

CREATE TABLE Students AS USE ds = "university" EXECUTE SELECT * from Students;

USE ... SHOW

USE ds = "university" SHOW TABLES;

USE … INSERT

USE ds = "company" INSERT INTO Employees VALUES ("Kevin", "DevOps", 110);

USE ds = "company" INSERT INTO Employees SELECT name, profession, 80 FROM Graduates;

USE ds = "company" INSERT INTO Employees USE ds = "university" EXECUTE SELECT name, subject, 2 * salary FROM Professors;

SYNC

SYNC INTO Graduates FROM SELECT * FROM Team WHERE salary < 90;

SYNC INTO Graduates USE ds = "university" EXECUTE SELECT name, subject, 80 FROM Students;

SYNC INTO Graduates ON(name) FROM SELECT * FROM Team;

SYNC INTO AnalyzeBudget FROM ANALYZE name, job, salary, avg(win.salary) AS avgSalary, median(win.salary) AS medianSalary, salary - median(win.salary) AS diff FROM Team WINDOW win ON true;

SYNC INTO PrepareBudget PREPARE name, job, salary, stack(salary), norm(salary) FROM Team ORDER BY salary ASC PARTITION BY job;

DELETE FROM

DELETE FROM Graduates;

DELETE FROM OpenJobs WHERE id == 1;

DELETE FROM Team WHERE job == "Sales" LIMIT 2;

DELETE FROM Team WHERE job == "Programming" LIMIT 2,1;

DELETE FROM OpenJobs JOIN BonusJobs ON OpenJobs.id == BonusJobs.id;

DELETE FROM Team JOIN OpenJobs ON Team.job == OpenJobs.job;

DELETE FROM SoftwareTeam LEFT ONLY JOIN BonusJobs ON SoftwareTeam.job == BonusJobs.job;

TRUNCATE

TRUNCATE TABLE SoftwareTeam;

DROP

DROP TABLE Graduates, BonusJobs;
DROP TABLE IF EXISTS Team;

Historical Database Query Syntax

Create a Historical Table

//create a PARTITION Column on TradeDate
CREATE PUBLIC TABLE HistoricalOrders(TradeDate LONG PARTITION, TradeId int, Symbol String, Price double, Quantity int) use PersistEngine="HISTORICAL";

Likewise, you could declare a BITMAP, FLAT, VARSIZE column in a similar manner. The general format is <DATA_TYPE> + <STORAGE_TYPE>:

//create a BITMAP2 STRING Column on Symbol
CREATE PUBLIC TABLE HistoricalOrders(TradeDate LONG PARTITION, TradeId int, Symbol String BITMAP2, Price double, Quantity int) use PersistEngine="HISTORICAL";

//create a VARSIZE3 STRING Column on Symbol
CREATE PUBLIC TABLE HistoricalOrders(TradeDate LONG PARTITION, TradeId int, Symbol String VARSIZE3, Price double, Quantity int) use PersistEngine="HISTORICAL";

//create a FLAT_NONULL Integer Column on Quantity
CREATE PUBLIC TABLE HistoricalOrders(TradeDate LONG PARTITION, TradeId int, Symbol String, Price double, Quantity int FLAT_NONULL) use PersistEngine="HISTORICAL";

Optimize procedure

AMIHDB has a unique optimization procedure that helps optimize the storage of data in the historical table and writes it to disk. To call this procedure, the syntax is as follows:

CALL __OPTIMIZE_HISTORICAL_TABLE("HistoricalOrders");

Diagnose table clause

In order for the end user to have a detailed view of a particular historical table, we can use the following syntax to diagnose the table:

DIAGNOSE TABLE HistoricalOrders;

CREATE INDEX Clause

//create a sort index on Quantity column
CREATE INDEX sortQuantityIndex on HistoricalOrders(Quantity);