Database

Database

Schema

Database and database owner (DBO)

Tablespace

Database

User

User

Role

Group/Role

Table

Table

Temporary tables

Temporary tables

Cluster

N/A

Column-level check constraint

Column-level check constraint

Column default

Column default

Unique key

Unique key or identity property for a column

Primary key

Primary key

Foreign key

Foreign key

Index

Non-unique index

PL/SQL Procedure

Transact-SQL (T-SQL) stored procedure

PL/SQL Function

T-SQL stored procedure

Packages

N/A

AFTER triggers

Triggers

BEFORE triggers

Complex rules

Triggers for each row

N/A

Synonyms

N/A

Sequences

Identity property for a column

Snapshot

N/A

View

View

User-defined data types make the data definition language code and procedural SQL code less portable across different database servers.

INTEGER

Four-byte integer, 31 bits, and a sign. May be abbreviated as 'INT' (this abbreviation was required prior to version 5).

NUMBER(10)

It is possible to place a table constraint on columns of this type (as an option) to force values between -2^31 and2^31. Or, place appropriate constraints such as: STATE_NO between 1 and 50

SMALLINT

Two-byte integer, 15 bits, and a sign.

NUMBER(6)

It is possible to place a table constraint on columns of this type (optionally) to force values between -2^15 and 2^15. Or, place appropriate constraints such as: STATE_NO between 1 and 50

TINYINT

One byte integer, 8 bits and no sign. Holds whole numbers between 0 and 255.

NUMBER(3)

You may add a check constraint of (x between 0 and 255) where x is column name.

REAL

Floating point number. Storage is four bytes and has a binary precision of 24 bits, a 7-digit precision.

Data can range from -3.40E+38 to 3.40E+38.

In Sybase the range of values and the actual representation is platform dependent. This can result in incorrect interpretation if data is moved between platforms. REAL numbers are stored in 4 bytes and can represent about 6 decimal digits with reasonable accuracy. Sybase REALs are mapped to the ANSI equivalent in Oracle.

FLOAT

The ANSI data type conversion to Oracle for REAL is FLOAT(63). By default, the Oracle Migration Workbench maps REAL to FLOAT(24) that stores up to 8 significant decimal digits in Oracle.

The Oracle NUMBER data type is used to store both fixed and floating-point numbers in a format that is compatible with decimal arithmetic. You may want to add a check constraint to constrain range of values. Also, you get different answers when performing operations on this data type as the Oracle NUMBER type is more precise and portable than REAL. Floating-point numbers can be specified in Oracle in the following format: FLOAT[(b)]. Where [(b)] is the binary precision b and can range from 1 to 126. [(b)] defaults to 126. To check what a particular binary precision is in terms of decimal precision, multiply [(b)] by 0.30103 and round up to the next whole number.

FLOAT

A floating point number. This column has 15-digit precision.

FLOAT

The ANSI data type conversion to Oracle for FLOAT(p) is FLOAT(p). The ANSI data type conversion to Oracle for DOUBLE PRECISION is FLOAT(126). By default, the Oracle Migration Workbench maps FLOAT to FLOAT(53), that stores up to 16 significant decimal digits in Oracle.

The Oracle NUMBER data type is used to store both fixed and floating-point numbers in a format compatible with decimal arithmetic.You get different answers when performing operations on this type due to the fact that the Oracle NUMBER type is much more precise and portable than FLOAT, but it does not have the same range. The NUMBER data type data can range from -9.99.99E+125 to 9.99.99E+125 (38 nines followed by 88 zeros).

NOTE: If you try to migrate floating point data greater than or equal to 1.0E+126 then Migration Workbench will fail to insert this data in the Oracle database and1 will return an error.This also applies to negative values less than or equal to -1.0E+126.

Floating-point numbers can be specified in Oracle using FLOAT[(b)], where [(b)] is the binary precision [(b)] and can range from 1 to 126. [(b)] defaults to 126.To check what a particular binary precision is in terms of decimal precision multiply [(b)] by 0.30103 and round up to the next whole number.

If they are outside of the range, large floating-point numbers will overflow, and small floating-point numbers will underflow.

BIT

A Boolean 0 or 1 stored as one bit of a byte. Up to 8-bit columns from a table may be stored in a single byte, even if not contiguous. Bit data cannot be NULL, except for Microsoft SQL Server 7.0, where null is allowed by the BIT data type.

NUMBER(1)

In Oracle, a bit is stored in a number(1) (or char). In Oracle, it is possible to store bits in a char or varchar field (packed) and supply PL/SQL functions to set / unset / retrieve / query on them.

CHAR(n)

Fixed-length string of exactly n 8-bit characters, blank padded. Synonym for CHARACTER.
0 < n < 256 for Microsoft SQL Server and Sybase Adaptive Server.
0 < n < 8000 for Microsoft SQL Server 7.0.

CHAR(n)

Pro*C client programs must use mode=ansi to have characters interpreted correctly for string comparison, mode=oracle otherwise.

VARCHAR(n)

Varying-length character string. 0 < n < 256 for Microsoft SQL Server and Sybase Adaptive Server.
0 < n < 8000 for Microsoft SQL Server 7.0.

VARCHAR2(n)

TEXT

Character string of 8-bit bytes allocated in increments of 2k pages. 'TEXT' is stored as a linked-list of 2024-byte pages, blank padded. TEXT columns can hold up to (231-1) characters.

CLOB

The CLOB field can hold up to 4GB.

IMAGE

Binary string of 8-bit bytes. Holds up to (231-1) bytes of binary data.

BLOB

The BLOB field can hold up to 4GB.

BINARY(n)

Fixed length binary string of exactly n 8-bit bytes.
0 < n < 256 for Microsoft SQL Server and Sybase Adaptive Server.
0 < n < 8000 for Microsoft SQL Server 7.0.

RAW(n)/BLOB

VARBINARY(n)

Varying length binary string of up to n 8-bit bytes.
0 < n < 256 for Microsoft SQL Server and Sybase Adaptive Server.
0 < n < 8000 for Microsoft SQL Server 7.0.

RAW(n)/BLOB

DATETIME

Date and time are stored as two 4-byte integers. The date portion is represented as a count of the number of days offset from a baseline date (1/1/1900) and is stored in the first integer. Permitted values are legal dates between 1st January, 1753 AD and 31st December, 9999 AD. Permitted values in the time portion are legal times in the range 0 to 25920000. Accuracy is to the nearest 3.33 milliseconds with rounding downward. Columns of type DATETIME have a default value of 1/1/1900.

DATE

The precision of DATE in Oracle and DATETIME in Microsoft SQL Server and Sybase Adaptive Server is different. The DATETIME data type has higher precision than the DATE data type. This may have some implications if the DATETIME column is supposed to be UNIQUE. In Microsoft SQL Server and Sybase Adaptive Server, the column of type DATETIME can contain UNIQUE values because the DATETIME precision in Microsoft SQL Server and Sybase Adaptive Server is to the hundredth of a second. In Oracle, however, these values may not be UNIQUE as the date precision is to the second. You can replace a DATETIME column with two columns, one with data type DATE and another with a sequence, in order to get the UNIQUE combination. It is preferable to store hundredths of seconds in the second column.

The Oracle TIMESTAMP data type can also be used. It has a precision of 1/10000000th of a second.

SMALL-DATETIME

Date and time stored as two 2-byte integers. Date ranges from 1/1/1900 to 6/6/2079. Time is the count of the number of minutes since midnight.

DATE

With optional check constraint to validate the smaller range.

MONEY

A monetary value represented as an integer portion and a decimal fraction, and stored as two 4-byte integers. Accuracy is to the nearest 1/10,000. When inputting Data of this type it should be preceded by a dollar sign ($). In the absence of the '$' sign, Microsoft SQL Server and Sybase Adaptive Server create the value as a float.

Monetary data values can range from -922,337,203,685,477.5808 to 922,337,203,685,477.5807, with accuracy to a ten-thousandth of a monetary unit. Storage size is 8 bytes.

NUMBER(19,4)

Microsoft SQL Server and Sybase Adaptive Server input MONEY data types as a numeric data type with a preceding dollar sign ($) as in the following example, select * from table_x where y > $5.00 You must remove the '$' sign from queries. Oracle is more general and works in international environments where the use of the '$' sign cannot be assumed. Support for other currency symbols and ISO standards through NLS is available in Oracle.

NCHAR(n)

Fixed-length character data type which uses the UNICODE UCS-2 character set. n must be a value in the range 1 to 4000. SQL Server storage size is two times n.

Note: Microsoft SQL Server storage size is two times n. The Oracle Migration Workbench maps columns sizes using byte semantics, and the size of Microsoft SQL Server NCHAR data types appear in the Oracle Migration Workbench Source Model with 'Size' specifying the number of bytes, as opposed to the number of Unicode characters. Thus, a SQL Server column NCHAR(1000) will appear in the Source Model as NCHAR(2000).

CHAR(n*2)

NVARCHAR(n)

Fixed-length character data type which uses the UNICODE UCS-2 character set. n must be a value in the range 1 to 4000. SQL Server storage size is two times n.

Note: Microsoft SQL Server storage size is two times n. The Oracle Migration Workbench maps columns sizes using byte semantics, and the size of Microsoft SQL Server NCHAR data types appear in the Oracle Migration Workbench Source Model with 'Size' specifying the number of bytes, as opposed to the number of Unicode characters. Thus, a SQL Server column NCHAR(1000) will appear in the Source Model as NCHAR(2000).

VARCHAR(n*2)

SMALLMONEY

Same as MONEY above except monetary data values from -214,748.3648 to +214,748.3647, with accuracy to one ten-thousandth of a monetary unit. Storage size is 4 bytes.

NUMBER(10,4)

Since the range is -214,748.3648 to 214,748.364, NUMBER(10,4) suffices for this field.

TIMESTAMP

TIMESTAMP is defined as VARBINARY(8) with NULL allowed. Every time a row containing a TIMESTAMP column is updated or inserted, the TIMESTAMP column is automatically incremented by the system. A TIMESTAMP column may not be updated by users.

NUMBER

You must place triggers on columns of this type to maintain them. In Oracle you can have multiple triggers of the same type without having to integrate them all into one big trigger. You may want to supply triggers to prevent updates of this column to enforce full compatibility.

SYSNAME

VARCHAR(30) in Microsoft SQL Server and Sybase Adaptive Server.

NVARCHAR(128) in Microsoft SQL Server 7.0.

VARCHAR2(30) and VARCHAR2(128) respectively.

CHAR

VARCHAR

NCHAR

NVARCHAR

BINARY

VARBINARY

DATETIME, SMALLDATETIME

DATETIMN

FLOAT

FLOATN

INT, SMALLINT, TINYINT

INTN

DECIMAL

DECIMALN

NUMERIC

NUMERICN

MONEY, SMALLMONEY

MONEYN

The Oracle Migration Workbench Source Model will display table system data types for each column.

Database Devices:

A database device is mapped to the specified physical disk files.

Datafiles:

One or more datafiles are created for each tablespace to physically store the data of all logical structures in a tablespace. The combined size of the datafiles in a tablespace is the total storage capacity of the tablespace. The combined storage capacity of a the tablespaces in a database is the total storage capacity of the database. Once created, a datafile cannot change in size. This limitation does not exist in Oracle.

Page:

Many pages constitute a database device. Each page contains a certain number of bytes.

Data Block:

One data block corresponds to a specific number of bytes, of physical database space, on the disk. The size of the data block can be specified when creating the database. A database uses and allocates free database space in Oracle data blocks.

Extent:

Eight pages make one extent. Space is allocated to all the databases in increments of one extent at a time.

Extent:

An extent is a specific number of contiguous data blocks, obtained in a single allocation.

N/A

Segments:

A segment is a set of extents allocated for a certain logical structure. The extents of a segment may or may not be contiguous on disk, and may or may not span the datafiles.

Segments (corresponds to OracleTablespace):

A segment is the name given to one or more database devices. Segment names are used in CREATE TABLE and CREATE INDEX constructs to place these objects on specific database devices. Segments can be extended to include additional devices as and when needed by using the SP_EXTENDSEGMENT system procedure.

The following segments are created along with the database:

• System segment
  Stores the system tables.
  
• Log segment
  Stores the transaction log.
  
• Default segment
  All other database objects are stored on this segment unless specified otherwise.

Segments are subsets of database devices.

Tablespace (corresponds to Microsoft SQL Server and Sybase Adaptive Server Segments):

A database is divided into logical storage units called tablespaces. A tablespace is used to group related logical structures together. A database typically has one system tablespace and one or more user tablespaces.

Tablespace Extent:

An extent is a specific number of contiguous data blocks within the same tablespace.

Tablespace Segments:

A segment is a set of extents allocated for a certain logical database object. All the segments assigned to one object must be in the same tablespace. The segments get the extents allocated to them as and when needed.

There are four different types of segments as follows:

• Data segment
  Each table has a data segment. All of the table's data is stored in the extents of its data segments. The tables in Oracle can be stored as clusters as well. A cluster is a group of two or more tables that are stored together. Each cluster has a data segment. The data of every table in the cluster is stored in the cluster's data segment.

Tablespace Segments (Cont):

• Index segment
  Each index has an index segment that stores all of its data.
  
• Rollback segment
  One or more rollback segments are created by the DBA for a database to temporarily store 'undo' information. This is the information about all the transactions that are not yet committed. This information is used to generate read-consistent database information during database recovery to rollback uncommitted transactions for users.

• Temporary segment
  Temporary segments are created by Oracle when a SQL statement needs a temporary work area to complete execution. When the statement finishes execution, the extents in the temporary segment are returned to the system for future use.

Log Devices:

These are logical devices assigned to store the log. The database device to store the logs can be specified while creating the database.

Redo Log Files:

Each database has a set of two or more redo log files. All changes made to the database are recorded in the redo log. Redo log files are critical in protecting a database against failures. Oracle allows mirrored redo log files so that two or more copies of these files can be maintained. This protects the redo log files against failure of the hardware the log file reside on.

Database Devices:

A database device contains the database objects. A logical device does not necessarily refer to any particular physical disk or file in the file system.

The database and logs are stored on database devices. Each database device must be initialized before being used for database storage. Initialization of the database device initializes the device for storage and registers the device with the server. After initialization, the device can be:

• Allocated to the free space available to a database
  
• Allocated to store specific user objects
  
• Used to store the transaction log of a database
  
• Labeled as default device to create and alter database objects

The SP_HELPDEVICES system procedure displays all the devices that are registered with the server. Use the DROP DEVICE DEVICE_NAME command to drop the device. The system administrator (SA) should restart the server after dropping the device.

A device can be labeled as a default device so that the new databases need not specify the device at the time of creation. Use the SP_DISKDEFAULT system procedure to label the device as a default device.

N/A

Dump Devices

These are logical devices. A database dump is stored on these devices. The DUMP DATABASE command uses the dump device to dump the database.

N/A

N/A

Control Files:

Each database has a control file. This file records the physical structure of the database. It contains the following information:

• database name
  
• names and locations of a database's datafiles and redo log files
  
• time stamp of database creation

It is possible to have mirrored control files. Each time an instance of an Oracle database is started, its control file is used to identify the database, the physical structure of the data, and the redo log files that must be opened for the database operation to proceed. The control file is also used for recovery if necessary. The control files hold information similar to the master database in Microsoft SQL Server and Sybase Adaptive Server.

Syntax:

Syntax:

Description:

A default database is assigned to each user. This database is made current when the user logs on to the server. A user executes the USE DATABASE_NAME command to switch to another database.

Syntax:

Syntax:

Description:

DISTINCT eliminates the duplicate rows.

The INTO clause and the items that follow it in the command syntax are optional, because Microsoft SQL Server and Sybase Adaptive Server allow SELECT statements without FROM clauses as can be seen in the following example:

SELECT...INTO allows you to insert the results of the SELECT statement into a table.

SELECT_LIST can contain a SELECT statement in the place of a column specification as follows:

The above example also shows the format for the column alias.

COMPUTE attaches computed values at the end of the query. These are called row_aggregates.

Outer joins are implemented as follows:

Description:

DISTINCT eliminates the duplicate rows.

The INSERT INTO <table> SELECT FROM.... construct allows you to insert the results of the SELECT statement into a table.

COLUMN ALIAS is defined by putting the alias directly after the selected COLUMN.

If you use TABLE ALIAS, the TABLE must always be referenced using the ALIAS.

You can also retrieve data from SYNONYMS.

EXPRESSION could be a column name, a literal, a mathematical computation, a function, several functions combined, or one of several PSEUDO-COLUMNS.
Outer joins are implemented as follows:

WHERE tab1.col1 = tab2.col1 (+);

Where all values from TAB1 are returned even if TAB2 does not have a match or

WHERE tab1.col1 (+) = tab2.col1;

where all values from TAB2 are returned even if TAB1 does not have a match.

If a GROUP BY clause is used, all non-aggregate select columns must be in a GROUP BY clause.

The FOR UPDATE clause locks the rows selected by the query. Other users cannot lock these row until you end the transaction. This clause is not a direct equivalent of the FOR BROWSE mode in Microsoft SQL Server and Sybase Adaptive Server.

where all values from TAB1 are returned even if TAB2 does not have a match or

where all values from TAB2 are returned even if TAB1 does not have a match.

If a GROUP BY clause is used, all non-aggregate select columns are needed.

FOR BROWSE keywords are used to get into browse mode. This mode supports the ability to perform updates while viewing data in an OLTP environment. It is used in front-end applications using DB-Library and a host programming language. Data consistency is maintained using the TIMESTAMP field in a multi-user environment. The selected rows are not locked; other users can view the same rows during the transaction. A user can update a row if the TIMESTAMP for the row is unchanged since the time of selection.

1993

1

1.3

1993

2

1.4

1993

3

3

1993

4

2.3

Microsoft SQL Server and Sybase Adaptive Server also support Oracle-style column aliases.

Syntax:

See the SELECT Statement section.

Syntax:

See the SELECT Statement section.

Description:

Non-aggregate SELECT columns must be in a GROUP BY clause.

Description:

All non-aggregate SELECT columns must be in a GROUP BY clause.

Syntax:

Syntax:

Description:

INTO is optional.

Inserts are allowed in a view provided only one of the base tables is undergoing change.

Description:

INTO is required.

Inserts can only be done on single table views.

Oracle lets you create functions that directly match most Microsoft SQL Server and Sybase Adaptive Server functions.

Syntax:

Syntax:

Description:

The FROM clause is used to get the data from one or more tables into the table that is being updated or to qualify the rows that are being updated.

Updates through multi-table views can modify only columns in one of the underlying tables.

Description:

A single subquery may be used to update a set of columns. This subquery must select the same number of columns (with compatible data types) as are used in the list of columns in the SET clause.

The CURRENT OF cursor clause causes the UPDATE statement to effect only the single row currently in the cursor as a result of the last FETCH. The cursor SELECT statement must have included in the FOR UPDATE clause.

Updates can only be done on single table views.

Syntax:

Syntax:

Description:

The first FROM in DELETE FROM is optional.

The second FROM clause is an Microsoft SQL Server and Sybase Adaptive Server extension that allows the user to make deletions based on the data in other tables. A subquery in the WHERE clause serves the same purpose.

Deletes can only be performed through single table views.

Description:

FROM is optional.

ALIAS can be specified for the table name as a correlation name, which can be used in the condition.

Deletes can only be performed through single table views

Equal to

=

Not equal to

!=

<>

^=

Less than

<

Greater than

>

Less than or equal to

<=

!>

Greater than or equal to

>=

!<

Greater than or equal to x and less than or equal to y

BETWEEN x AND y

Less than x or greater than y

NOT BETWEEN x AND y

Pattern Matches

a followed by 0 or more characters

a followed by exactly 1 character

a followed by any character between x and z

a followed by any character except those between x and z

a followed by %

LIKE 'a%'

LIKE 'a_'

LIKE'a[x-z]'

LIKE'a[^x-z]'

LIKE 'a%'

ESCAPE '

Does not match pattern

NOT LIKE

No value exists

IS NULL

A value exists

IS NOT NULL

At least one row returned by query

EXISTS (query)

No rows returned by query

NOT EXISTS (query)

Equal to a member of set

IN =ANY

= SOME

Not equal to a member of set

NOT IN != ANY <> ANY

!= SOME <> SOME

Less than a member of set

< ANY

< SOME

Greater than a member of set

> ANY

> SOME

Less than or equal to a member of set

<= ANY

!> ANY

<= SOME

Greater than or equal to a member of set

>= ANY

!< ANY

>= SOME

Equal to every member of set

=ALL

Not equal to every member of set

!= ALL <> ALL

Less than every member of set

< ALL

Greater than every member of set

> ALL

Less than or equal to every member of set

<= ALL

!> ALL

Greater than or equal to every member of set

>= ALL

!< ALL

Add

+

Subtract

-

Multiply

*

Divide

/

Modulo

v

%

mod(x, y)

Concatenate

s

+

||

Identify Literal

'this is a string'

'this is also a string'

Distinct row from either query

UNION

All rows from both queries

UNION ALL

All distinct rows in both queries

d

INTERSECT

All distinct rows in the first query but not in the second query

d

MINUS

bit and

&

bit or

|

bit exclusive or

^

bit not

~

ascii(char)

ascii(char)

Returns the ASCII equivalent of the character.

char(integer_expression)

chr(integer_expression)

Converts the decimal code for an ASCII character to the corresponding character.

charindex(specified_exp, char_string)

instr(specified_exp, char_string, 1, 1)

Returns the position where the specified_exp first occurs in the char_string.

convert(data type, expression, [format])

to_char, to_number, to_date, to_label, chartorowid, rowtochar, hextochar, chartohex

Converts one data type to another using the optional format. The majority of the functionality can be matched. Refer to Oracle9iSQL Reference, Release 1 (9.0.1) for more information.

datalength(expression)

g

Computes the length allocated to an expression, giving the result in bytes.

difference(character_exp, character_exp)

d

Returns the numeric difference of the SOUNDEX values of the two strings.

isnull(variable, new_value)

nvl(variable, new_value)

If the value of the variable is NULL, the new_value is returned.

lower(char_exp)

lower(char_exp)

Converts uppercase characters to lowercase characters.

ltrim(char_exp)

ltrim(char_exp)

Truncates trailing spaces from the left end of char_exp.

patindex(pattern,

column_name)

Returns the position of the pattern in the column value. The pattern can have wild characters. This function also works on TEXT and BINARY data types.

replicate(char_exp, n)

rpad(char_exp, length(char_exp)*n, ')

Produces a string with char_exp repeated n times.

reverse(char_string)

Reverses the given char_string.

right(char_exp, n)

substr(char_exp, (length(char_exp)

Returns the part of the string starting at the position given by n from the right and extending up to the end of the string.

rtrim(char_exp)

rtrim(char_exp)

Truncates the trailing spaces from the right end of char_exp.

soundex(exp)

soundex(exp)

Returns phonetically similar expressions to the specified exp.

space(int_exp)

rpad(' ', int_exp-1, ')

Generates a string with int_exp spaces.

str(float_exp, length)

to_char(float_exp)stuff(char_exp, start, length, replace_str)substr(char_exp, 1, start) ||replace_str ||substr(char_exp, start+length)

Replaces a substring within char_exp with replace_str.

substring(char_exp, start, length)

Works on IMAGE and TEXT data types

substr(char_exp, start, length)

Does not work with LONG and LONG_RAW data types

Replaces a substring within char_exp with replace_str.

textptr(column_name)

d

Returns a pointer as a varbinary(16) data type for a named IMAGE or TEXT column.

textvalid('column_name', text_pointer)

h

Returns 1 if the specified text_pointer is valid for the specified column_name. The column must be of type TEXT or IMAGE.

upper(char_exp)

upper(char_exp)

Converts lowercase characters to uppercase characters.

datalength(expression)

lengthb

Computes the length allocated to an expression, giving the result in bytes.

isnull(variable, new_value)

nvl(variable, new_value)

If the value of the variable is NULL, the new_value is returned.

The above functions tables list all the Microsoft SQL Server and Sybase Adaptive Server character manipulation functions. They do not list all the Oracle functions. There are many more Oracle character manipulation functions that you can use.

Adds the int_exp number of days to the date contained in datetime_var.

Adds the int_exp number of months to the date contained in datetime_var.

Adds the int_exp number of years to the date contained in datetime_var.

Returns the difference between the dates specified by the datetime1 and datetime2 variables. This difference is calculated in the number of days.

Returns the difference between the dates specified by the datetime1 and datetime2 variables. This difference is calculated in the number of months.

Returns the difference between the dates specified by the datetime1 and datetime2 variables. This difference is calculated in the number of years.

Returns the specified part of the date as an integer. The Microsoft SQL Server and Sybase Adaptive Server DATETIME has a higher precision than Oracle DATE. For this reason, it is not always possible to find an equivalent format string in Oracle to match the datepart in Microsoft SQL Server and Sybase Adaptive Server. See the Data Types section of this chapter for more information about conversion of the DATETIME data type.

Returns the specified part of the date as a character string (name). The Microsoft SQL Server and Sybase Adaptive Server DATETIME has a higher precision than Oracle DATE'. For this reason, it is not always possible to find an equivalent format string in Oracle to match the datepart in Microsoft SQL Server and Sybase Adaptive Server.

Returns the system date.

abs(n)

abs(n)

acos(n)

acos(n)

asin(n)

atan(n)

atan(n)

atn2(n,m)

ceiling(n)

ceil(n)

cos(n)

cos(n)

cot(n)

degrees(n)

exp(n)

exp(n)

floor(n)

floor(n)

log(n)

ln(n)

log10(n)

log(base,number)

pi()

power(m,n)

power(m,n)

radians(n)

rand(n)

round(n[,m])

round(n[,m])

sign(n)

sign(n)

sin(n)

sin(n)

sqrt(n)

sqrt(n)

tan(n)

tan(n)

Microsoft SQL Server and Sybase Adaptive Server locking is fully automatic and does not require intervention by users.

Microsoft SQL Server and Sybase Adaptive Server apply exclusive locks for INSERT, UPDATE, and DELETE operations. When an exclusive lock is set, no other transaction can obtain any type of lock on those objects until the original lock is in place.

For non-update or read operations, a shared lock is applied. If a shared lock is applied to a table or a page, other transactions can also obtain a shared lock on that table or page. However, no transaction can obtain an exclusive lock. Therefore, Microsoft SQL Server and Sybase Adaptive Server reads block the modifications to the data.

Update locks:

Update locks are held at the page level. They are placed during the initial stages of an update operation when the pages are being read. Update locks can co-exist with shared locks. If the pages are changed later, the update locks are escalated to exclusive locks.

Oracle locking is fully automatic and does not require intervention by users. Oracle features the following categories of locks:

Data locks (DML locks) to protect data.The 'table locks' lock the entire table and 'row locks' lock individual rows.

Dictionary locks (DDL locks) to protect the structure of objects.

Internal locks to protect internal structures, such as files.

DML operations can acquire data locks at two different levels; one for specific rows and one for entire tables.

Row-level locks:

An exclusive lock is acquired for an individual row on behalf of a transaction when the row is modified by a DML statement. If a transaction obtains a row level lock, it also acquires a table (dictionary) lock for the corresponding table. This prevents conflicting DDL (DROP TABLE, ALTER TABLE) operations that would override data modifications in an on-going transaction.

Intent locks:

Microsoft SQL Server and Sybase Adaptive Server locking is fully automatic and does not require intervention by users. Microsoft SQL Server and Sybase Adaptive Server apply exclusive locks for INSERT, UPDATE, and DELETE operations. When an exclusive lock is set, no other transaction can obtain any type of lock on those objects until the original lock is in place. For non-update or read operations, a shared lock is applied. If a shared lock is applied to a table or a page, other transactions can also obtain a shared lock on that table or page. However, no transaction can obtain an exclusive lock. Therefore, Microsoft SQL Server and Sybase Adaptive Server reads block the modifications to the data.

Extent locks:

Extent locks lock a group of eight database pages while they are being allocated or freed. These locks are held during a CREATE or DROP statement, or during an INSERT that requires new data or index pages.

A list of active locks for the current server can be seen with SP_LOCK system procedure.

Table-level data locks can be held in any of the following modes:

Row share table lock (RW):

This indicates that the transaction holding the lock on the table has locked rows in the table and intends to update them. This prevents other transactions from obtaining exclusive write access to the same table by using the LOCK TABLE table IN EXCLUSIVE MODE statement. Apart from that, all the queries, inserts, deletes, and updates are allowed in that table.

Row exclusive table lock (RX):

This generally indicates that the transaction holding the lock has made one or more updates to the rows in the table. Queries, inserts, deletes, updates are allowed in that table.

Share lock (SL):

Share row exclusive lock(SRX)

Exclusive lock (X):

The dynamic performance table V$LOCK keeps the information about locks.

Microsoft SQL Server and Sybase Adaptive Server do not have a row-level locking feature.

Microsoft SQL Server and Sybase Adaptive Server apply a page-level lock, which effectively locks all rows on the page, whenever any row in the page is being updated. This is an exclusive lock whenever the data is being changed by DML statements.

Microsoft SQL Server 7.0 implements a form of row-level locking.

Microsoft SQL Server 7.0 escalates locks at row level to page level automatically.

SELECT statements are blocked by exclusive locks that lock an entire page.

Oracle has a row-locking feature. Only one row is locked when a DML statement is changing the row.

Microsoft SQL Server and Sybase Adaptive Server provide the HOLDLOCK function for transaction-level read consistency. Specifying a SELECT with HOLDLOCK puts a shared lock on the data. More than one user can execute a SELECT with HOLDLOCK at the same time without blocking each other.

When one of the users tries to update the selected data, HOLDLOCK blocks the update until the other users commit, rollback, or attempt an update and a deadlock occurs. This means that HOLDLOCK prevents other transactions from updating the same data until the current transaction is in effect.

Read consistency as supported by Oracle does the following:

• Ensures that the set of data seen by a statement is consistent at a single point-in-time and does not change during statement execution
  
• Ensures that reads of database data do not wait for other reads or writes of the same data
  
• Ensures that writes of database data do not wait for reads of the same data
  
• Ensures that writes wait for other writes only if they attempt to update identical rows in concurrent transactions

To provide read consistency, Oracle creates a read-consistent set of data when a table is being read and simultaneously updated.

Read consistency functions as follows:

1. When an update occurs, the original datavalues changed by the update are recorde in rollback segments.
   
2. While the update remains part of an uncommitted transaction, any user that reads the modified data views the original data values. Only statements that start afteranother user's transaction is committed reflect the changes made by the transaction.

You can specify that a transaction be read only using the following command:

After completion, any statement not within a transaction is automatically committed.A statement can be a batch containing multiple T-SQL statements that are sent to the server as one stream. The changes to the database are automatically committed after the batch executes. A ROLLBACK TRAN statement subsequently executed has no effect. In Microsoft SQL Server and Sybase Adaptive Server, transactions are not implicit. Start logical transaction with a BEGIN TRANSACTION clause. Logical transactions can be committed or rolled back as follows.

Use COMMIT TRAN to commit the transaction to the database. You have the option to specify the transaction name. Use ROLLBACK TRAN to roll back the transaction. You can set savepoints to roll back to a certain point in the logical transaction using the following command:

Roll back to the specified SAVEPOINT with the following command:

Statements are not automatically committed to the database. The COMMIT WORK statement is required to commit the pending changes to the database.

Oracle transactions are implicit. This means that the logical transaction starts as soon as data changes in the database.

COMMIT WORK commits the pending changes to the database.

ROLLBACK undoes all the transactions after the last COMMIT WORK statement.

Savepoints can be set in transactions with the following command:

The following command rolls back to the specified SAVEPOINT:

Two-phase commit is automatic and transparent in Oracle. Two-phase commit operations are needed only for transactions which modify data on two or more databases.

Microsoft SQL Server and Sybase Adaptive Server allow you to nest BEGIN TRAN/COMMIT TRAN statements. When nested, the outermost pair of transactions creates and commits the transaction. The inner pairs keep track of the nesting levels. A ROLLBACK command in the nested transactions rolls back to the outermost BEGIN TRAN level, if it does not include the name of the SAVEPOINT. Most Microsoft SQL Server and Sybase Adaptive Server applications require two-phase commit, even on a single server. To see if the server is prepared to commit the transaction, use PREPARE TRAN in two-phase commit applications.

Completed transactions are written to the database device at CHECKPOINT. A CHECKPOINT writes all dirty pages to the disk devices since the last CHECKPOINT.

Calls to remote procedures are executed independently of any transaction in which they are included.

When a CHECKPOINT occurs, the completed transactions are written to the database device. A CHECKPOINT writes all dirty pages to the disk devices that have been modified since last checkpoint