set

enables and disables advanced aggregation at the session level.

allows multiple variable assignments in one statement. The set @variable = expression command is an identical—and an alternative—command to select @variable = expression in Transact-SQL.

expression includes constant, function, any combination of constants, and functions connected by arithmetic or bitwise operators, or a subquery.

determines how Adaptive Server behaves when an arithmetic error occurs. The two arithabort options, arithabort arith_overflow and arithabort numeric_truncation, handle different types of arithmetic errors. You can set each option independently or set both options with a single set arithabort on or set arithabort off statement.

• arithabort arith_overflow – specifies the Adaptive Server behavior following a divide-by-zero error, range overflow during an explicit or implicit datatype converson, or a domain error. This type of error is serious. The default setting, arithabort arith_overflow on, rolls back the entire transaction in which the error occurs. If the error occurs in a batch that does not contain a transaction, arithabort arith_overflow on does not roll back earlier commands in the batch; however, Adaptive Server does not execute any statements in the batch that follow the error-generating statement.

  Setting arith_overflow to on refers to the execution time, not to the level of normalization to which Adaptive Server is set.

  If you set arithabort arith_overflow off, Adaptive Server aborts the statement that causes the error, but continues to process other statements in the transaction or batch.

• arithabort numeric_truncation – specifies the Adaptive Server behavior following a loss of scale by an exact numeric type during an implicit datatype conversion. (When an explicit conversion results in a loss of scale, the results are truncated without warning.) The default setting, arithabort numeric_truncation on, aborts the statement that causes the error, but Adaptive Server continues to process other statements in the transaction or batch. If you set arithabort numeric_truncation off, Adaptive Server truncates the query results and continues processing.

determines whether Adaptive Server displays a message after a divide-by-zero error or a loss of precision. By default, the arithignore option is set to off. This causes Adaptive Server to display a warning message after any query that results in numeric overflow. To have Adaptive Server ignore overflow errors, use set arithignore on. You can omit the optional arith_overflow keyword without any effect.

establishes the number of rows that are buffered in local server memory before being transferred using the bulk copy interface.

Use this option only with CIS for transferring rows to a remote server using select into.

View your current setting using the @@bulkarraysize global variable.

number – indicates the number of rows to buffer. If the rows being transferred contain text, unitext , image or java ADTs, then the bulk copy interface ignores the current setting for array size and uses a value of 1. Also, the array size actually used will never exceed the value of @@bulkbatchzise. If @@bulkbatchsize is smaller than array size, then the smaller value is used.

The initial value of the array size is inherited by new connections from the current setting of the configuration property cis bulk insert array size, which defaults to 50. Setting this value to 0 will reset the value to the default.

establishes the number of rows transferred to a remote server via select into proxy_table when the bulk interface is used. The bulk interface is available to all Adaptive Servers, as well as DirectConnect for Oracle version 12.5.1.

Use this option only with CIS for transferring rows to a remote server using select into.

View your current setting using the @@bulkbatchsize global variable.

The bulk interface allows a commit after a specified number of rows. This allows the remote server to free any log space being consumed by the bulk transfer operation, and enables the transfer of large data sets from one server to another without filling the transaction log.

The initial value of the batch size is inherited by new connections from the current setting of the configuration property cis bulk insert batch size, which by default is 0. A value of 0 indicates that no rows should be committed until after the last row is transferred.

If a string is given as an argument in place of the chronological value the server interprets it as a datetime value regardless of its apparent precision. This is the default behavior and indicated by a builtin_date_strings value of 0.

Changing the builtin_date_strings value to 1 makes the server interpret the argument strings as bigdatetimes. This will affect the result of chronological builtins.

• chained – begins a transaction just before the first data retrieval or data modification statement at the beginning of a session and after a transaction ends. In chained mode, Adaptive Server implicitly executes a begin transaction command before the following statements: delete, fetch, insert, lock table, open, select, and update. You cannot execute set chained within a transaction.

• close on endtran – causes Adaptive Server to close all cursors opened within a transaction at the end of that transaction. A transaction ends by the use of either the commit or rollback statement. However, only cursors declared within the scope that sets this option (stored procedure, trigger, and so on) are affected. For more information about cursor scopes, see the Transact-SQL User’s Guide.

  For more information about the evaluated configuration, see the System Administration Guide.

• nocount – controls the display of rows affected by a statement. set nocount on disables the display of rows; set nocount off reenables the count of rows.

• noexec – compiles each query but does not execute it. noexec is often used with showplan. After you set noexec on, no subsequent commands are executed (including other set commands) until you set noexec off.

• parseonly – checks the syntax of each query and returns any error messages without compiling or executing the query. Do not use parseonly inside a stored procedure or trigger.

• self_recursion – determines whether Adaptive Server allows triggers to cause themselves to fire again (this is called self recursion). By default, Adaptive Server does not allow self recursion in triggers. You can turn this option on only for the duration of a current client session; its effect is limited by the scope of the trigger that sets it. For example, if the trigger that sets self_recursion on returns or causes another trigger to fire, this option reverts to off. This option works only within a trigger and has no effect on user sessions.

• showplan – generates a description of the processing plan for the query. The results of showplan are of use in performance diagnostics. showplan does not print results when it is used inside a stored procedure or trigger. For parallel queries, showplan output also includes the adjusted query plan at runtime, if applicable. See the Performance and Tuning Guide.

• sort_resources – generates a description of the sorting plan for a create index statement. The results of sort_resources are of use in determining whether a sort operation is done serially or in parallel. When sort_resouces is on, Adaptive Server prints the sorting plan but does not execute the create index statement. See Chapter 24, “Parallel Sorting,” in the Performance and Tuning Guide.

enables or disables character set conversion between Adaptive Server and a client. If the client is using Open Client DB-Library release 4.6 or later, and the client and server use different character sets, conversion is turned on during the login process and is set to a default based on the character set the client is using. You can also use set char_convert charset to start conversion between the server character set and a different client character set.

charset can be either the character set’s ID or a name from syscharsets with a type value of less than 2000.

set char_convert off turns conversion off so that characters are sent and received unchanged. set char_convert on turns conversion on if it is turned off. If character set conversion was not turned on during the login process or by the set char_convert command, set char_convert on generates an error message.

If you request character set conversion with set char_convert charset, and Adaptive Server cannot perform the requested conversion, the conversion state remains the same as it was before the request. For example, if conversion is set to off prior to the set char_convert charset command, conversion remains turned off if the request fails.

When the with no_error option is included, Adaptive Server does not notify an application when characters from Adaptive Server cannot be converted to the client’s character set. Error reporting is initially turned on when a client connects with Adaptive Server: if you do not want error reporting, you must turn it off for each session with set char_convert {on | charset} with no_error. To turn error reporting back on within a session, use set char_convert {on | charset} with error.

Whether or not error reporting is turned on, the bytes that cannot be converted are replaced with ASCII question marks (?).

See the System Administration Guide for a more complete discussion of error handling in character set conversion.

makes CIS the remote procedure call (RPC) handling mechanism the default mechanism for Shared Disk Cluster (SDC) handling

assigns names to the client.

• clientname client_name – assigns a client an individual name. This is useful for differentiating among clients in a system where many clients connect to Adaptive Server using the same client name. After you assign a new name to a user, they appear in the sysprocesses table under the new name.

  client_name is the new name you assign to the user.

• clienthostname host_name – assigns a host an individual name. This is useful for differentiating among clients in a system where many clients connect to Adaptive Server using the same host name. After you assign a new name to a host, it appears in the sysprocesses table under the new name.

  host_name is the new name you assign to the host.

• clientapplname application_name – assigns an application an individual name. This is useful for differentiating among clients in a system where many clients connect to Adaptive Server using the same application name. After you assign a new name to an application, it appears in the sysprocesses table under the new name.

  application_name is the new name you assign to the application.

enables or disables compression for the session:

• on – enables data compression on new data for tables and partitions on which compression is configured. Adaptive Server compresses all rows that qualify.

• off – inserts and updates all new data as uncompressed. Inserts that trigger a page compression ignore the uncompressed rows. Updated rows remain uncompressed. Uncompressed rows that are compressed during an update (rows inserted as uncompressed remain uncompressed until explicitly compressed).

• default – resets the compression level to the default setting (inserts and updates are compressed according to the table or partition setting).

  Unlike most set parameters, if you execute set export_options before issuing set compression in a nested procedure, Adaptive Server does not export the compression level to the parent procedure’s context.

causes Adaptive Server to return the number of rows for each cursor fetch request from a client application. The number can be a numeric literal with no decimal point or a local variable of type integer. If the number is less than or equal to zero, the value is set to 1. You can set the cursor rows option for a cursor, whether it is open or closed. However, this option does not affect a fetch request containing an into clause. cursor_name specifies the cursor for which to set the number of rows returned.

specifies the following settings:

• datefirst number – uses numeric settings to specify the first day of the week. The us_english language default is Sunday. To set the first day of the week, use the following:

  To set the first day of the week as	Use this setting
  Monday	1
  Tuesday	2
  Wednesday	3
  Thursday	4
  Friday	5
  Saturday	6
  Sunday (us_english language default)	7
  Regardless of which day you set as the first day of the week, the value of that first day becomes 1. This value is not the same as the numeric setting you use in set datefirst n. For example, if you set Sunday as your first day of the week, its value is 1. If you set Monday as your first day of the week, Monday’s value becomes 1. If you set Wednesday as your first day of the week, Wednesday’s value becomes 1, and so on.

• dateformat format – sets the order of the date parts month/day/year for entering datetime, smalldatetime, date or time data. Valid arguments are mdy, dmy, ymd, ydm, myd, and dym. The us_english language default is mdy.

• language language – is the official name of the language that displays system messages. The language must be installed on Adaptive Server. The default is us_english.

sets deferred name resolution on for the current session only.

determines when log records are written to disk. With the delayed_commit parameter set to true, the log records are asynchronously written to the disk and control is returned to the client without waiting for the IO to complete.

The session-level setting overrides any existing the database-level setting. Change delayed_commit to its default to revert back to the database-level setting.

determines the amount of logging for insert, update, and delete (DML) operations. Valid values are:

• minimal – Adaptive Server attempts to log no changes for the DML statements. In most cases, Adaptive Server generates little or no logging to syslogs.

• default – Adaptive Server disables session-specific minimal logging and uses the logging mode you enabled for individual tables based on table-specific and database-wide logging levels.

creates the encryption key’s password to encrypt or decrypt data on an insert, update, delete, select, alter table, or select into statement.

• password_phrase – is the explicit password specified with the create encryption key or alter encryption key command to protect the key.

• key – indicates that Adaptive Server uses this password to decrypt the key when accessing any column encrypted by the named key

• keyname – may be supplied as a fully qualified name. For example:

  [[database.][owner].]keyname
  

• column – specifies that Adaptive Server use this password only in the context of encrypting or decrypting the named column. End users do not necessarily know the name of the key that encrypts a given column.

• column_name – name of the column on which you are setting an encryption password. Supply column_name as:

  [[ database.][ owner ]. ]table_name.column_name
  

Adaptive Server’s default behavior is to reset any set parameter changes that are set by a trigger or system procedure after they finish running. Enabling set export_options allows you to retain the session settings that are set by a system procedure or trigger for the duration of the session.

For example, this enables set export_options:

set export_options on

This disables set export_options and returns Adaptive Server to the default behavior:

set export_options off

determines whether Adaptive Server displays a warning message when Transact-SQL extensions to entry-level ANSI SQL are used. By default, Adaptive Server does not tell you when you use nonstandard SQL. This option does not disable SQL extensions. Processing completes when you issue the non-ANSI SQL command.

determines when Adaptive Server returns messages to the user. By default, messages are stored in a buffer until the query that generated them is completed or the buffer is filled to capacity. Use set flushmessage on to return messages to the user immediately, as they are generated.

captures plans in stored procedures without actually executing them.

causes the query optimizer to use the order of the tables in the from clause of a query as the join order for the query plan. forceplan is generally used when the optimizer fails to choose a good plan. Forcing an incorrect plan can have severely bad effects on I/O and performance. See the Performance and Tuning Guide.

1> set forceplan on
2> select * from table1, table2 
    where table2.id *= table1.id

determines whether explicit inserts into a table’s IDENTITY column are allowed. (Updates to an IDENTITY column are never allowed.) This option can be used only with base tables. It cannot be used with views or set within a trigger.

After setting identity_insert on for the table, the table owner, database owner, or system administrator can manually insert any legal value greater than 5. For example, inserting a value of 55 would create a large gap in IDENTITY column values:

insert stores_cal 
 (syb_identity, stor_id, stor_name)
values (55, "5025", "Good Reads")
select syb_identity from stores_cal

id_col  
-------  
      1 
      5 
     55 
 
 (3 rows affected)

If identity_insert is then set to off, Adaptive Server assigns an IDENTITY column value of 55 + 1, or 56, for the next insertion. If the transaction that contains the insert statement is rolled back, Adaptive Server discards the value 56 and uses a value of 57 for the next insertion.

Unless you have created a unique index on the IDENTITY column, Adaptive Server does not verify the uniqueness of the inserted value; you can insert any positive integer.

The table owner, database owner, or system administrator can use the set identity_insert table_name on command on a table with an IDENTITY column to enable the manual insertion of a value into an IDENTITY column. However, only the following users can actually insert a value into an IDENTITY column, when identity_insert is on:

• Table owner

• Database owner:

  • If granted explicit insert permission on the column by the table owner

  • Impersonating the table owner by using the setuser command

Setting identity_insert table_name off restores the default behavior by prohibiting explicit inserts to IDENTITY columns. At any time, you can use set identity_insert table_name on for a single database table within a session.

With set identity_update on, you can explicitly update the value of the IDENTITY column on a table. identity_update changes the identity column value for the qualified rows. When identity_update is enabled, you can update the identity value to any value greater than 0. However, if the input value is greater than the identity burn max value, a new set of ID values is allocated, and the identity burn max value on the OAM page is updated accordingly. If update is included in a transaction, the new identity burn max value cannot be rolled back. You can use syb_identity to point to the identity column for update. For example:

update table_name set syb_identity = value
where clause

Adaptive Server does not check for duplicates entries or verify that entries are unique. You can update an existing value to any positive integer within the range allowed by the column's declared precision. You can check for duplicate entries by creating a unique index on the identity column

when enabled, sets limits the scan of a table with an or clause.

Index unions (also known as an or strategy) are used for queries that contain or clauses. For example:

select * from titleauthor where au_id = "409-56-7008" or title_id = "PC8888" 

If index_union is:

• Enabled – this example uses an index on au_id to find the row IDs (RIDs) of all titleauthor tuples with au_id = "409-56-7008", and uses an index on title_id to find the RIDs of all titleauthor tuples with title_id = "PC8888". Adaptive Server then performs a union on all RIDs to eliminate duplicates. The resulting RIDs are joined with a RidJoin to access the data tuples.

• Disabled – Adaptive Server does not use an index union strategy in a query to limit the table scan. Instead, it uses other access paths on the table (in the example above, it would use a table scan for table titleauthor), and applies the or clause as a filter in the scan operator.

is on by default. If the server-level setting for literal_autoparam is on, this option enables and disables use of that feature. If the server level setting is off, this setting has no effect.

specifies the settings for a lock.

• wait – specifies the length of time that a command waits to acquire locks before aborting and returning an error.

• numsecs – specifies the number of seconds a command is to wait to acquire a lock. Valid values are from 0 to 2147483647, the maximum value for an integer.

• lock nowait – specifies that if a command cannot acquire a lock immediately, it returns an error and fails. set lock nowait is equivalent to set lock wait 0.

enables the capture of QP metrics at the session level, set the capture to “on”. Query processing (QP) metrics identify and compare empirical metric values in query execution. When a query is executed, it is associated with a set of defined metrics that are the basis for comparison in QP metrics.

returns the position of specified keywords (with relation to the beginning of the query) in Transact-SQL statements. The keyword list is a comma-separated list that can include any of the following Transact-SQL constructs: select, from, order, compute, table, procedure, statement, param, and execute. Adaptive Server returns offsets if there are no errors.

This option is used in Open Client DB-Library only.

generates diagnostics output in text format.

The valid values for show_option are:

• show – shows the basic syntax common to all modules

• show_lop – shows the logical operators (scans, joins, etc.) used

• show_managers – shows data structure managers used during optimization.

• show_log_props – shows the logical properties (row count, selectivity, etc.) evaluated.

• show_parallel – shows details of parallel query optimization

• show_histograms – shows the processing of histograms associated with SARG/Join columns

• show_abstract_plan – shows the details of an abstract plan

• show_search_engine – shows the details of the join ordering algorithm

• show_counters – shows the optimization counters

• show_best_plan – shows the details of the best query plan selected by the optimizer

• show_pio_costing – shows estimates of physical input/output (reads/writes from/to the disk)

• show_lio_costing – shows estimates of logical input/output (reads/writes from/to memory)

• show_elimination – shows partition elimination

• show_missing_stats – shows details of useful statistics missing from SARG/Join columns

See Chapter 4, “Displaying Query Optimization Strategies And Estimates,” in Query Optimizer for more information.

sets the timeout limit for the optimizer. The valid range of values for opttimeoutlimit 0 to 4000 ms, with 0 indicating no optimization limit.

specifies an upper limit for the number of worker processes used in the parallel execution of a query. This number must be less than or equal to the number of worker processes per query, as set by the max parallel degree configuration parameter. The @@parallel_degree global variable stores the current setting.

introduces an abstract plan command.

• dump – enables or disables capturing abstract plans for the current connection. If a group_name is not specified, the plans are stored in the default group, ap_stdout.

• load – enables or disables loading abstract plans for the current connection. If a group_name is not specified, the plans are loaded from the default group, ap_stdin.

• group_name – is the name of the abstract plan group to use for loading or storing plans.

See Chapter 30, “Creating and Using Abstract Plans,” in the Performance and Tuning Guide.

when used with set plan load, stores hash keys for up to 20 queries from an abstract plan group in a per-user cache.

generates an XML document for the diagnostic output. The valid values for show are:

• show_exec_xml – gets the compiled plan output in XML, showing each of the query plan operators.

• show_execio_xml – gets the plan output along with estimated and actual IOs. This also includes the query text.

• show_opt_xml – gets optimizer diagnostic output, which shows all of the different components like logical operators, output from the managers, some of the search engine diagnostics, and the best query plan.

• show_lop_xml – gets the output logical operator tree in XML.

• show_managers_xml – shows the output of the different component managers during the preparation phase of the query optimizer.

• show_log_props_xml – shows the logical properties for a given equivalence class (one or more group of relations in the query).

• show_parallel_xml – shows the diagnostics related to the optimizer while generating parallel query plans.

• show_histograms_xml – shows diagnostics related to histograms and the merging of histograms.

• show_abstract_plan_xml – shows the AP generation/application.

• show_search_engine_xml – shows the search engine related diagnostics.

• show_counters_xml – shows plan object construction/destruction counters.

• show_best_plan_xml – shows the best plan in XML.

• show_pio_costing_xml – shows actual PIO costing in XML.

• show_lio_costing_xml – shows actual LIO costing in XML.

• show_elimination_xml – shows partition elimination in XML.

• client – when specified, output goes to the client.

• message – when specified, output goes to an internal message buffer.

See Chapter 4, “Displaying Query Optimization Strategies And Estimates,” in Query Processor for more information.

sets the optimization goal.

• allrows_mix – is the default optmization goal, and the most useful goal in a mixed-query environment. It balances the needs of OLTP and DSS query environments.

• allrows_dss – is the most useful goal for operational DSS queries of medium to high complexity. Currently, this goal is provided on an experimental basis.

See Chapter 1, “Understanding Query Processing in Adaptive Server” in Query Processor for more information about optimization plans.

sets the timeout at the session level, where n is any integer between 0 and 1000. See Chapter 1, “Understanding Query Processing in Adaptive Server” in Query Processor for more information about optimization plans.

enables or disables replacing existing abstract plans during plan capture mode. By default, plan replacement is off.

enables or disables large I/Os to the data cache.

generates trace information to session output, reporting on the statement for which the minimally logged mode of a table was overridden by some other rules, such as presence of referential integrity constraints, deferred mode choice, name of the table affected and a description of the affecting rules, and so on.

controls sending of output parameters that a stored procedure generates back to the client. set proc_output_params off suppresses sending the output parameters back to the client. The default for this parameter is on.

controls sending of a return status TDS token back to the client. set proc_return_status off suppresses sending the return status token to the client, and isql client does not display the (return status = 0) message. The default for this parameter is on.

specifies whether Adaptive Server executes parallel queries when an insufficient number of worker processes is available. Under these circumstances, if:

• process_limit_action is set to quiet, Adaptive Server silently adjusts the plan to use a degree of parallelism that does not exceed the number of available processes.

• process_limit_action is set to warning when an insufficient number of worker processes are available, Adaptive Server issues a warning message when adjusting the plan

• process_limit_action is set to abort, Adaptive Server aborts the query and issues an explanatory message an insufficient number of worker processes are available.

determines whether Adaptive Server recognizes delimited identifiers within double quotation marks. By default, quoted_identifier is off and all identifiers must either:

• Conform to the rules for valid identifiers.

• Be enclosed in brackets.

If you use set quoted_identifier on, double quotation marks behave the way brackets do, and you can use table, view, and column names that begin with a nonalphabetic character, including characters that would not otherwise be allowed, or are reserved words, by enclosing the identifiers within double quotation marks. Delimited identifiers cannot exceed 28 bytes, may not be recognized by all front-end products, and may produce unexpected results when used as parameters to system procedures.

When quoted_identifier is on, all character strings enclosed within double quotes are treated as identifiers. Use single quotes around character or binary strings.

is the maximum degree to which any intermediate data stream is re-partitioned for semantic purposes. See Chapter 2, “Parallel Query Processing” in Query Processor for more information about setting the value of max repartition degree for a session.

sets the SQL replication threshold at the session level. If set repthreshold is invoked in a stored procedure, its scope is that of the procedure. If it is invoked in a user session, its scope is that of the session.

A user can alter the scope of the threshold and set the session threshold using a login trigger, in which case you need not explicitly set the session threshold at login.

For example,

create proc myproc
as
	set repthreshold 777
---------------------

alter login sa modify login script 'myproc'
----------------------
option changed. 

(Return status = 0

The procedure ’myproc’ is invoked each time the user sa logs in, and ensures that the replication threshold is set at 777 for the whole session.

Another way to alter the scope of the threshold is to use set export_options:

create proc p2
as
	set repthreshold222
	set export_options on
---------------------------

After p2 is executed the threshold remains set at 222.The hierarchy of thresholds is:

Session > Table > Database

When you set the session threshold to 0, the only replication thresholds that still matter are table and database thresholds, if they have been specified, in that order. If you set no threshold, the threshold defaults to 50 rows.

To set the table level threshold, see sp_setrepdbmode; to set the database level threshold, see sp_setrepdefmode, both in the Reference Manual: Procedures.

The session threshold is exportable; a stored procedure can set the threshold and turn the export options setting ON. Adaptive Server enforces the new threshold in the invoking procedure or session.

overrides the global value max resource granularity and sets it to a session specific value, which influences whether Adaptive Server uses memory-intensive operation or not. See Chapter 2, “Parallel Query Processing” in Query Processor for more information.

turns the specified role on or off during the current session. When you log in, all system roles that have been granted to you are turned on.

Use set role role_name off to turn a role off, and set role role_name on to turn it back on again, as needed.

• "sa_role" | "sso_role" | "oper_role" are system roles. If you are not a user in the current database, and if there is no “guest” user, you cannot set sa_role off, because there is no server user ID for you to assume.

• role_name – is the name of any user-defined role created by the system security officer. User-defined roles are not turned on by default. To set user-defined roles to activate at login, the user or the system security officer must use set role on.

• with passwd – specifies the password to activate the role. If a user-defined role has an attached password, you must specify the password to activate the role.

causes Adaptive Server to stop processing the query (select, insert, update, or delete) after the specified number of rows are affected. The number can be a numeric literal with no decimal point or a local variable of type integer. To turn this option off, use:

set rowcount 0

You can determine the current value for set rowcount with the @@setrowcount global variable. For example:

select @@setrowcount
____________
           37

specifies the maximum session-specific degree of parallelism for hash-based scans (parallel index scans and parallel table scans on nonpartitioned tables). This number must be less than or equal to the current value of the max scan parallel degree configuration parameter. The @@scan_parallel_degree global variable stores the current setting.

specifies whether Adaptive Server sends the LOB or the locator that references the LOB in a result set sent to the client. When the option is off (the default), Adaptive Server sends the LOB.

is identical to set proxy, with this exception: set session authorization follows the SQL standard, while set proxy is a Transact-SQL extension.

generates additional information when a command executes.

• on – generates extra diagnostics about the logging mode for a DML statement. Displays the selected logging mode for the current statement and session, and the user executing the DML.

• off – disables show_exec_info.

allows you to print the SQL text for ad-hoc queries, stored procedures, cursors, and dynamic prepared statements.

You do not need to enable the set show_sqltext before you execute the query (as you do with commands like set showplan on) to collect diagnostic information for a SQL session. Instead you can enable it while the commands are running to help determine which query is performing poorly and diagnose their problems.

Before you enable show_sqltext, you must first enable dbcc traceon to display the output to standard out:

dbcc traceon(3604)

The syntax for set show_sqltext is:

set show_sqltext {on | off}

For example, this enables show_sqltext:

set show_sqltext on

Once set show_sqltext is enabled, Adaptive Server prints all SQL text to standard out for each command or system procedure you enter. Depending on the command or system procedure you run, this output can be extensive.

To disable show_sqltext, enter:

set show_sqltext off

is on by default. If the server-level setting for statement_cache is on, this option enables and disables use of that feature. If the server level setting is off, this setting has no effect.

displays various types of statistics information

• io – displays statistics for each table referenced in the statement:

  • The number of times the table is accessed (scan count)

  • The number of logical reads (pages accessed in memory)

  • The number of physical reads (database device accesses)

  For each command, statistics io displays the number of buffers written.

  If Adaptive Server has been configured to enforce resource limits, statistics io also displays the total I/O cost.

• subquerycache – displays the number of cache hits, misses, and the number of rows in the subquery cache for each subquery.

• time – displays the amount of time Adaptive Server used to parse and compile for each command. For each step of the command, statistics time displays the amount of time Adaptive Server used to execute the command. Times are given in milliseconds and timeticks, the exact value of which is machine-dependent.

• plancost – displays the query statistics in a tree format.

  When you enable set statistics plancost, Adaptive Server abbreviates the names for lio, pio, and row to l, p, and r, respectively.

• simulate – specifies that the optimizer should use simulated statistics to optimize the query.

See Chapter 34, “Using the set statistics Commands” in the Performance and Tuning Guide.

determines whether the server propagates transactions to servers that do not support Adaptive Server transaction coordination services. The default value is inherited from the value of the strict dtm enforcement configuration parameter.

determines whether Adaptive Server raises a SQLSTATE exception when an insert or update command truncates a char, unichar, varchar or univarchar string. If the truncated characters consist only of spaces, no exception is raised. The default setting, off, does not raise the SQLSTATE exception, and the character string is silently truncated.

(clusters only) specifies the system view for a session, and controls the materialization of fake tables, which impact the output of stored procedures such as sp_who.

• instance – sets the system view for the current instance.

• cluster – sets the system view for the cluster.

• clear – clears any session-level setting, reverting to the system_view setting for the logical cluster hosting that spid. Enter select @@system_view to check the current value.

allows you to set trace flags and switch names locally and server-wide.

• serverwide – optional and will set a switch serverwide ON or OFF. The default is session-specific.

• on – trace flags are switched on.

• off – trace flags are switched off.

• trace_flag – a sequence of numbers (the old traceflag numbers) and/or switch names.

• option – an optional sequence of switch options. Valid values are:

  • override – this option is necessary to enable a non-documented switch names or trace flags

  • no_info – this option is used to surpress any informational warnings

specifies the maximum size in bytes of text, unitext, or image type data that is returned with a select statement. The @@textsize global variable stores the current setting.

The default setting is 32K in isql. Some client software sets other default values. To reset textsize to the default size (32K), use:

set textsize 0

once enabled, saves all SQL text for the current session to the specified file, each SQL text batch appending to the previous batch.

The syntax to enable tracing is:

set tracefile file_name [off] [for spid

The syntax to disable tracing is:

set tracefile off [for spid]

Where:

• file_name – is the full path to the file in which you are saving the SQL text. If you do not specify a directory path, Adaptive Server creates the file in $SYBASE.

  If file_name contains special characters (“:”, “/”, and so on) other than numbers and letters, you must include file_name in quotes. For example, this file_name must be in quotes because of the “/” for the directory structure:

  set tracefile '/tmp/mytracefile.txt' for 25
  

  If file_name does not contain special characters and you want to save it to $SYBASE, it does not require quotes. For example, this file_name does not need to be in quotes:

  set tracefile mytracefile.txt
  

• off – disables the tracing for this session or spid.

• spid – server process ID whose SQL text you want saved to a trace file. Only the users with the SA or systems security officer role can enable tracing for other spids. You cannot save the SQL text for system tasks (such as the housekeeper or the port manager).

sets the transaction isolation level for your session. After you set this option, any current or future transactions operate at that isolation level.

• read uncommitted | 0 – scans at isolation level 0 do not acquire any locks. Therefore, the result set of a level 0 scan may change while the scan is in progress. If the scan position is lost due to changes in the underlying table, a unique index is required to restart the scan. In the absence of a unique index, the scan may be aborted.

  By default, a unique index is required for a level 0 scan on a table that does not reside in a read-only database. You can override this requirement by forcing the Adaptive Server to choose a nonunique index or a table scan, as follows:

  select * from table_name (index table_name)
  

  Activity on the underlying table may cause the scan to be aborted before completion.

• read committed | 1 – by default, the Adaptive Server transaction isolation level is read committed or 1, which allows shared read locks on data.

• repeatable read | 2 – prevents nonrepeatable reads.

• serializable | 3 – specifies isolation level 3, Adaptive Server applies a holdlock to all select and readtext operations in a transaction, which holds the queries’ read locks until the end of that transaction. If you also set chained mode, that isolation level remains in effect for any data retrieval or modification statement that implicitly begins a transaction.

controls the handling of remote procedure calls. If this option is set to on, when a transaction is pending, the RPC is coordinated by Adaptive Server. If this option is set to off, the remote procedure call is handled by the Adaptive Server site handler. The default value is inherited from the value of the enable xact coordination configuration parameter.