Options

DEBUG [boolean]

DBPATH [path]

LOGPATH [path]

REPOPATH [path]

LANG [language]

FAIRLOCK [boolean]

• By default, non-fair locking is used. Read transactions are favored, and transactions that no not access databases can be evaluated even if the limit for parallel transactions (specified via PARALLEL) is reached. This prevents update operations from blocking all other requests. For example, the DBA can still be used to see which jobs are running even if the queue is full.
• When fair locking is enabled, read and write transactions are treated equally (first in, first out). This avoids starvation of updates and should be used when prompt evaluation of updates is critical.

CACHETIMEOUT [seconds]

WRITESTORE [boolean]

HOST [host]

PORT [port]

SERVERPORT [port]

USER [name]

• The default value is overwritten if a client specifies its own credentials.
• If the default value is empty, login will only be possible if the client specifies credentials.
• The option can also be changed on the command line with -U.

PASSWORD [password]

• The default value is overwritten if a client specifies its own credentials.
• If the default value is empty, authentication will only be possible if the client supplies credentials.
• The option can also be changed on the command line with -P.
• Please note that it is a security risk to specify your password in plain text.

AUTHMETHOD [method]

• If basic access is chosen, the client can still request digest authentication.
• This is different for digest access, which cannot be overwritten.
• With custom authentication, the server will not do any authentication.

SERVERHOST [host|ip]

PROXYHOST [host]

PROXYPORT [port]

NONPROXYHOSTS [hosts]

IGNORECERT [boolean]

TIMEOUT [seconds]

KEEPALIVE [seconds]

PARALLEL [number]

• If FAIRLOCK is enabled, the number of parallel transactions will never exceed the specified value.
• If the option is disabled (which is the default), the limit only applies to transactions that access databases.
• The main reason for allowing parallel operations is to prevent slow transactions from blocking all other operations. A higher number of parallel operations may increase disk activity and thus slow down queries. In some cases, a single transaction may even give you better results than any parallel activity.

LOG [string]

• data: database directory
• stdout: standard out
• stderr: standard error
• slf4j: SLF4J logger (supported levels: trace, debug, info, warn, error)

LOGEXCLUDE [string]

LOGCUT [string]

LOGMSGMAXLEN [length]

LOGTRACE [boolean]

WEBPATH [path]

• The option is ignored if BaseX is deployed as web servlet.
• It cannot be assigned via the web.xml file, as it is evaluated before the configuration files are parsed.

GZIP [boolean]

• The option can also be enabled on the command line with -g.
• It cannot be assigned via the web.xml file, as the option is evaluated before the configuration files are parsed.
• In addition to the Jetty defaults (GET requests, exclusion of binaries, MSIE 6.0, etc.), POST and PUT requests are supported.

RESTXQPATH [path]

PARSERESTXQ [PARSERESTXQ]

• If 0 is specified, the directory is parsed every time a RESTXQ function is called.
• A positive value defines the idle time, in seconds, after which parsing is enforced. The default value is 3: Changes in the RESTXQ directory is detected after 3 seconds without RESTXQ function calls.
• Monitoring is completely disabled if a negative value is specified.

See RESTXQ Preliminaries for more details.

RESTXQERRORS [RESTXQERRORS]

RESTPATH [path]

HTTPLOCAL [boolean]

STOPPORT [port]

• The listener for stopping the web server will only be started if the specified value is greater than 0.
• The option is ignored if BaseX is used as a Web Application or started via Maven.
• This option can also be changed when running the HTTP server on the command line with -s.

MAINMEM [boolean]

• Most queries are evaluated faster in main-memory mode, but all data is lost if the BaseX instance in which the database was created is shut down.
• It is not possible to store binary resources in a main-memory database.
• A main-memory database will have no disk representation. However, it is possible to export the database via the EXPORT command, and create a new database from the exported file in a second step.
• This option will not be available for db:create because the database would not be accessible anymore after database creation, i.e., outside the query scope.

ADDCACHE [boolean]

CREATEFILTER [filter]

ADDARCHIVES [boolean]

ARCHIVENAME [boolean]

SKIPCORRUPT [boolean]

ADDRAW [boolean]

PARSER [type]

CSVPARSER [options]

JSONPARSER [options]

HTMLPARSER [options]

STRIPWS [boolean]

• If the option is set to true, leading and trailing whitespace from text nodes is stripped and empty text nodes are discarded.
• The flag should be disabled if a document contains mixed content.
• The flag can also be enabled on the command line with -w.
• If the option is enabled, whitespace of an element and its descendants can locally be preserved with the xml:space="preserve" attribute:

<xml>
  <title>
    Demonstrating the STRIPWS flag
  </title>
  <text xml:space="preserve">To <b>be</b>, or not to <b>be</b>, …</text>
</xml>

If whitespace is stripped, indent=yes can be assigned to the SERIALIZER option to get properly indented XML output.

See BaseX 10: Whitespace for more information on whitespace handling.

STRIPNS [boolean]

INTPARSE [boolean]

• Performance: Documents, in particular small ones, are parsed faster.
• Entity support: More than 2000 HTML character references are decoded.
• Fault tolerance: Invalid characters are replaced with the Unicode replacement character FFFD (�).

By default, Java’s Xerces parser is used as it supports all features of the XML standard, including XINCLUDE, or recursive entity expansion if specified in DTDs.

DTD [boolean]

XINCLUDE [boolean]

CATALOG [path]

TEXTINDEX [boolean]

ATTRINDEX [boolean]

TOKENINDEX [boolean]

FTINDEX [boolean]

TEXTINCLUDE [names]

ATTRINCLUDE [names]

TOKENINCLUDE [names]

FTINCLUDE [names]

MAXLEN [int]

MAXCATS [int]

UPDINDEX [boolean]

• The current value of this option is assigned to new databases. It can be changed for existing databases by running OPTIMIZE with the ALL keyword or db:optimize and true() as second argument.
• After each update, the value indexes are refreshed as well. Incremental updates are currently not available for the full-text index and database statistics.
• Find more details in the article on Index Structures.

AUTOOPTIMIZE [boolean]

• With each update, outdated indexes and database statistics are recreated.
• As a result, the index structures will always be up-to-date.
• However, updates can take much longer, so this option should only be activated for medium-sized databases.
• The value of this option is assigned once to a new database. It can be reassigned by running OPTIMIZE or db:optimize.

SPLITSIZE [num]

• By default, if the value is set to 0, some heuristics are applied, based on the current memory consumption. Usually, this works fine.
• If explicit garbage collection is disabled when running Java (e.g. via the JVM option -XX:+DisableExplicitGC), you may need to choose a custom split size.
• You can e.g. start with 1000000 (one million) index operations and adjust this value in the next steps.
• The larger the assigned value is, the fewer splits take place, and main-memory consumption increases.

STEMMING [boolean]

CASESENS [boolean]

DIACRITICS [boolean]

LANGUAGE [lang]

STOPWORDS [path]

QUERYINFO [boolean]

MIXUPDATES [MIXUPDATES]

BINDINGS [vars]

• Variable names and values must be separated by the equals sign =.
• Multiple variables must be delimited by the comma sign ,.
• Commas in values must be duplicated.
• Variables may optionally be introduced with a leading dollar sign.
• If a variable uses a namespace different from the default namespace, it can be specified with the Expanded QNames notation.

This option can also be set on the command line with -b.

INLINELIMIT [INLINELIMIT]

• The XQuery compiler inlines functions to speed up query evaluation and to allow for further optimizations.
• A function body will be inlined if its number of expressions does not exceed the limit specified by this option.
• Function inlining can be disabled by setting the value to 0. This can be useful for debugging, as you will get a full stack trace when an error occurs.
• The limit can be locally overwritten via the %basex:inline annotation (follow the link to get more information on function inlining).

UNROLLLIMIT [UNROLLLIMIT]

• Loops with few iterations are unrolled by the XQuery compiler to enable further optimizations.
• If the limit is increased, more optimizations will take place, but the memory consumption and compile time will increase.
• See Loop Unrolling for more details.

ENFORCEINDEX [boolean]

COPYNODE [boolean]

let $a := <a/>
let $b := <b>{ $a }</b>
return $b/a is $a

This step can be expensive and memory consuming. If the option is disabled, child nodes will only be linked to the new parent nodes, and the upper query returns true. The option should be used carefully as it changes the semantics of XQuery. It should preferably be used in Pragmas.

TAILCALLS [TAILCALLS]

WITHDB [WITHDB]

• No locks are created for the two functions (see limitations of database locking for more details).
• Access to local and external resources is faster, as the database lookup is skipped.

DEFAULTDB [DEFAULTDB]

FORCECREATE [boolean]

CHECKSTRINGS [boolean]

WRAPJAVA [mode]

LSERROR [error]

RUNQUERY [boolean]

RUNS [num]

SERIALIZE [boolean]

SERIALIZER [params]

• Names and values are separated by the equals sign =.
• Multiple parameters are delimited by the comma sign ,.
• Commas must be duplicated if they appear as literals in values.

The option can also be set on the command line with -s.

EXPORTER [params]

• Names and values are separated by the equals sign =.
• Multiple parameters are delimited by the comma sign ,.
• Commas must be duplicated if they appear as literals in values.

XMLPLAN [boolean]

FULLPLAN [boolean]

AUTOFLUSH [boolean]

WRITEBACK [boolean]

• This option can also be toggled on the command line with -u.
• Please take in mind that no backup is created from your original files.
• The serialization options can be controlled via the EXPORTER option.

MAXSTAT [num]