Mark Logic Command: put

Puts a document onto a Mark Logic server.

Description

Puts a file (or expression) from the local system onto a Mark Logic server.

Options

-c,-connect uri	Use the connection string instead of $MLCONNECT
-uri uri	Use the specified uri as the uri of the destination document, Otherwise uses the base uri of the file or expression.
-baseuri base-uri	Use the specifed base uri instead of the base URI from the file or expression
-r,-recurse	Recursively put directories. If any argument is a directory then it is recursed
-m,-maxfiles num	Send files in batch at must num files in one transaction
-t,-text	Stores the document as "text" content type
-b,-binary	Stores the document as "binary" content type
-x,-xml	Stores the document as "xml" content type
-d,-mkdirs	Create directories as needed. Only creates direct parent directories of documents
-maxthreads n	Use at most n threads, default=1
-collection name [-collection name2 ...]	Put document(s) in the named collection(s)
-forest id [-forest id2 ...]	Put document(s) in the forest(s) specified by ID
-perm,-permission perm [-perm perm ...]	Apply the permission(s) to the document(s)
-repair none\|full\|default	Specify the repair mode
-buffer size	Set the transfer buffer size
-language lang	Set the language
-md5	Store an xmd5 property with the document
-namespace namespace	Set the document namespace
-resolve,+resolve	Resolve (do not Resolve) Entities before storing the document
-quality quality	Set the document quality
-locale locale	Set the locale
-f,-filenames file	Use a file (or stdin in "-") to specify the list of files to put
-delete	Deletes file after successfully putting to marklogic.
-stream port	Reads documents from the named port.
expression \| file	The file (or expression , variable, port or url) to put to the Mark Logic server.

If file is omitted then stdin is used.
If neither -x, -t or -b is specified then the content type is determined by the server.

If the uri contains {random} it will be replaced by a random number.
If the uri contains {seq} it will be replaced by the by an incrementing sequence number.
This allows you to completely ignore the local filename and put batches of files with unique names.
For example
ml:put -uri "/contacts/contact{seq}.xml" *.xml

will generate uris like

/contacts/contact1.xml
/contacts/contact2.xml

This is particularly useful if the input came from a tool like xsplit
where the filenames are not particulary useful but you dont want to do anything fancy like xmove

Examples

Assuming the prefix "ml" is imported as the marklogic module.

Put the file "test.xml" using its full path as the uri

ml:put test.xml

Puts the file "test.xml" using the uri "myfile.xml"

ml:put -uri myfile.xml test.xml

Puts the dynamic document <foo/> to the ML server using the uri "test.xml"

ml:put -uri test.xml <[ <foo/> ]>

Puts the result of an xquery (via stdin) to the Mark Logic server using the uri "test.xml"

xquery -q 'myquery.xquery' -i input.xml | ml:put -uri test.xml

Copies an entire directory tree "modules" to the /modules/ directory on the ML server, using at most 100 files per batch

ml:put -baseuri / -r -m 100 modules

Inserts a document into 2 collections

ml:put -uri test.xml -collection foo -collection bar test.xml

Permissions

Permissions can be specified with the -permission (or -perm) option. Permissions take the form of
*capability:role*

Capibility may be one of

u|update Update
r|read Read
i|insert Insert
x|execute Execute

Role is the role name

Example, add a document with read and update capability to the "user" role

ml:put -uri test.xml -perm read:user -perm u:user test.xml

If the -md5 option is set then a property is stored with the document of the form
<xmd5 md5="md5checksum" length="length" />

This can be used to implement an efficient "sync" command which only updates the destination if it has changed.

Notes

Like most xmlsh commands, anywhere an input file is expected, a file, uri, expression, port, or variable expression may be used instead.

For detailed explanation of all the options, see the MarkLogic XCC Developers Guide and the XCC Package Summary.

The put command uses the XCC Session.insertContent API.

Streaming

Careful use of the -filenames option can allow streaming multithreaded puts. The filenames are read and put (possibly batched by the -m and -maxthreads options) as they are read. If this comes from a pipe (like stdin) the list of files can be generated dynamically by a streaming process such as xsplit. This allows large xml files or sets of files to be streamed to a MarkLogic server as they are generated.

Note that if -mkdirs is specifed with -filenames then the list of files must be read in its entirety so that directories can be precreated before any files are put. This negates the ability to stream the puts as the filenames are read.

Port Streaming

If -stream port option is given then documements are read directly from the named pipe ( see CommandM=Xmkpipe xmkpipe ). Since the documents have no name, a base URI and uri template need to be supplied.
A URI template is a string containing {seq} or {random} which evaluates to a sequential or random number.
For example -uri test{seq}.xml will produce incremental unique URI's.

Example using xsplit to split a large XML file into smaller documents and put to marklogic without use of temporary files

import module ml=marklogic
xmkpipe -xml pipe

ml:put -baseuri /test/ -uri "doc{seq}.xml" -maxthreads 4 -maxfiles 100 -stream pipe &

xsplit -stream pipe file.xml 
xmkpipe -close pipe 
wait

MarkLogic Extension Module

xmlsh - XML Shell : MarkLogicPut