create table TransactionDetail (
      Id counter,
      Patient varchar(50),
      Item integer,
      ArdetItem integer,
      `Date` date,
      TransactionType integer,
      Description varchar(100),
      Acctno integer,
      Orthodontist integer,
      Office integer,
      Amount currency,
      DTD integer,
      MTD integer,
      constraint thePrimaryKey Primary Key(Id)
   )

   create table Application (
      ComputerName varchar(50), 
      Application varchar(100), 
      LoggedOn date, 
      LoggedOff date, 
      constraint Index1 primary key(ComputerName, Application)
   )

   create index ARLogPatient on ARLog(Patient)

   alter table rparty add Phone1 text(25)

   alter table patient drop address1
   alter table patient drop address2
   alter table patient drop city
   alter table patient drop state
   alter table patient drop zip

   drop table newdeposit
   drop table newsplit

   --widen the phone number for a table
   alter table rparty add newphone text(50)
   update rparty set newphone = phone
   alter table rparty drop phone
   alter table rparty add phone text(50)
   update rparty set phone=newphone
   alter table rparty drop newphone

alter table movie add genre2 text(50)
update movie set genre2 = genre
select genre2, {fn locate(',',genre2)}, left(genre2,{fn locate(',',genre2)}-1) from movie where {fn locate(',',genre2)} > 0
update movie set genre2 = left(genre2,{fn locate(',',genre2)}-1) where {fn locate(',',genre2)} > 0
select genre, subgenre from movie
update movie set genre2 = 'Sci-Fi' where genre2 like 'Sci%'
alter table movie add genre3 text(100)
update movie set genre3 = genre
update movie set genre = genre2
alter table drop genre2
alter table movie add subgenre text(100)
update movie set subgenre = genre3
alter table movie drop genre3
alter table movie drop genre2
alter table movie add temp text(100)
update movie set temp = subgenre
update movie set subgenre = left(subgenre,{fn locate(',',subgenre)}-1) where {fn locate(',',subgenre)} > 0
select subgenre, {fn substring( subgenre, {fn locate(',',subgenre)}+1 ,  100} from movie where {fn locate(',',subgenre)} > 0
select {fn substring('abcde',3,2)} from movie
update movie set subgenre = {fn substring( subgenre, {fn locate(',',subgenre)}+1 ,  100} where {fn locate(',',subgenre)} > 0
update movie set subgenre = {fn ltrim(subgenre)}

• Construct Primary Keys From Non-Duplicated Data
  If you can find a unique key such as the datatimestamp in combination with something else this is preferable to using sequences. The problem with sequences is that they eventually have gaps and/or wrap around. This means additional maintenance later. For example, using a datetime stamp combined with an IPAddress is guaranteed to be unique for all time.
• Create Meaningless Keys
  What I mean is don't put the office number into the key with the customer initials. Eventually, the customer will change locations and then your entire keyset becomes meaningless (because you can't trust it's meaning on any others).
• Allow For Unlimited Items
  If you're going to allow multiple phone number, don't stop at some arbitrary limit in a flat file.
• Use Words Instead of Codes
  Words like high, medium, low are superior to numeric constants such as 3, 2, 1 which must be interpreted by programmers and users.
• Use Meaningful Column Names
  Names like LastName, EnteredOn, LastBilledOn are superior to LN0001, ENT02, BD002. I know this seems obvious but I've worked for major software companies that couldn't see the rediculousness of the latter approach.
• Use Server Side Processing
  If you want blazing speed for your batch processes, create stored procedures for them. If your database doesn't support stored procedures (such as MySQL), create processes that run on the server for all batch operations. The disadvantage of stored procedures is having to maintain the code because it's now separate from the main application code.
• Avoid Triggers
  Only use triggers for debugging and tracking security issues. Remove them as soon as you've found what you're looking for. They seriously degrade system performance.