Part VI. Updating XML Documents. Marc H. Scholl (DBIS, Uni KN) XML and Databases Winter 2007/08 587

Size: px

Start display at page:

Download "Part VI. Updating XML Documents. Marc H. Scholl (DBIS, Uni KN) XML and Databases Winter 2007/08 587"

Hilda Dawson
5 years ago
Views:

1 Part VI Updating XML Documents Marc H. Scholl (DBIS, Uni KN) XML and Databases Winter 7/8 587

2 Outline of this part Updating XML Trees Update Specification XQuery Update Facility Impact on XPath Accelerator Encoding Impacts on Other Encoding Schemes ORDPATH Encoding What can Pathfinder/MonetDB do? Marc H. Scholl (DBIS, Uni KN) XML and Databases Winter 7/8 588

3 . Updating XML Trees Update Specification Updating XML trees Throughout the course, up to now, we have not been looking into updates to XML documents at all. 57 If we want to discuss efficiency/performance issues w.r.t. mappings of XML documents to databases, though, we need to take modifications into account as well as pure retrieval operations. As always during physical database design, there is a trade-off between accelerated retrieval and update performance. While there is a whole host of languages for querying (i.e., read access to) XML documents, there is not yet an update language (for write access) that has been agreed upon. We will briefly sketch the language primitives that have been discussed for expressing XML updates. The current way of expressing updates is via the XQuery Update Facility Marc H. Scholl (DBIS, Uni KN) XML and Databases Winter 7/8 589

4 . Updating XML Trees Update Specification Updates and tree structures During our discussion of XQuery, we have seen that tree construction has been a major concern. Updates, however, cannot be expressed with XQuery. Yet, we need to be able to specify modifications of existing XML documents/fragments as well. The basic necessary update functionality is largely agreed upon, syntax and semantic details, however, are subject to discussion. We certainly need to be able to express: modification of all aspects (name, attributes, attribute values, text contents) of XML nodes, and modifications of the tree structure (insert/delete/replace nodes or subtrees) and to rename and transform them. Like in the SQL case, target node(s) of such modifications should be identifiable by means of expressions in an/any XML query language. Marc H. Scholl (DBIS, Uni KN) XML and Databases Winter 7/8 59

5 . Updating XML Trees XQuery Update Facility XQuery Update Facility: New XQuery expressions XQuery expressions 58 ExprSingle ::= FLWORExpr QuantifiedExpr TypeswitchExpr IfExpr InsertExpr DeleteExpr RenameExpr ReplaceExpr TransformExpr OrExpr N.B. Updating expressions (insert, delete, rename, replace) lead to a loss of type/validation information at the affected nodes. Such information may be recovered by revalidation. 58 Syntax and examples taken from the web site. Marc H. Scholl (DBIS, Uni KN) XML and Databases Winter 7/8 59

6 . Updating XML Trees XQuery Update Facility Node insertion An insert expression is an updating expression that inserts copies of zero or more nodes into a designated position with respect to a target node. Syntax InsertExpr ::= "insert" ("node" "nodes") SourceExpr InsertExprTargetChoice TargetExpr InsertExprTargetChoice ::= (("as" ("first" "last"))? "into") "after" "before" SourceExpr ::= ExprSingle TargetExpr ::= ExprSingle Marc H. Scholl (DBIS, Uni KN) XML and Databases Winter 7/8 59

7 Node insertion: Examples. Updating XML Trees XQuery Update Facility Insert a year element after the publisher of the first book. insert node <year>5</year> after fn:doc("bib.xml")/books/book[]/publisher Navigating by means of several bound variables, insert a new police report into the list of police reports for a particular accident. insert node $new-police-report as last into fn:doc("insurance.xml")/policies /policy[id = $pid] /driver[license = $license] /accident[date = $accdate] /police-reports Marc H. Scholl (DBIS, Uni KN) XML and Databases Winter 7/8 59

8 . Updating XML Trees XQuery Update Facility Node deletion A delete expression deletes zero or more nodes from an XDM instance. The keywords node and nodes may be used interchangeably, regardless of how many nodes are actually deleted. Syntax DeleteExpr ::= "delete" ("node" "nodes") TargetExpr TargetExpr ::= ExprSingle Delete the last author of the first book in a given bibliography. delete node fn:doc("bib.xml")/books/book[]/author[last()] Delete all messages that are more than 65 days old. delete nodes / /message [fn:currentdate() - date > xs:daytimeduration("p65d")] Marc H. Scholl (DBIS, Uni KN) XML and Databases Winter 7/8 59

9 . Updating XML Trees XQuery Update Facility Node replacement Syntax ReplaceExpr ::= "replace" ("value" "of")? "node" TargetExpr "with" ExprSingle TargetExpr ::= ExprSingle Replace takes two forms, depending on whether value of is specified: If value of is not specified, a replace expression replaces one node with a new sequence of zero or more nodes. The replacement nodes occupy the position in the node hierarchy that was formerly occupied by the node that was replaced. Hence, an attribute node can be replaced only by zero or more attribute nodes, and an element, text, comment, or processing instruction node can be replaced only by zero or more element, text, comment, or processing instruction nodes. If value of is specified, a replace expression is used to modify the value of a node while preserving its node identity. Marc H. Scholl (DBIS, Uni KN) XML and Databases Winter 7/8 595

10 . Updating XML Trees XQuery Update Facility Node replacement: Examples Replace the publisher of the first book with the publisher of the second book. replace node fn:doc("bib.xml")/books/book[]/publisher with fn:doc("bib.xml")/books/book[]/publisher Increase the price of the first book by ten percent. replace value of node fn:doc("bib.xml")/books/book[]/price with fn:doc("bib.xml")/books/book[]/price *. Marc H. Scholl (DBIS, Uni KN) XML and Databases Winter 7/8 596

11 . Updating XML Trees XQuery Update Facility Renaming nodes A rename expression replaces the name property of a data model node with a new QName. Syntax RenameExpr ::= "rename" "node" TargetExpr "as" NewNameExpr Rename the first author element of the first book to principal-author. rename node fn:doc("bib.xml")/books/book[]/author[] as "principal-author" Rename the first author element of the first book to the QName that is the value of the variable $newname. rename node fn:doc("bib.xml")/books/book[]/author[] as $newname Marc H. Scholl (DBIS, Uni KN) XML and Databases Winter 7/8 597

12 Renaming is local!. Updating XML Trees XQuery Update Facility The effects of a rename expression are limited to its target node, descendants are not affected. Global change of names or namespaces needs explicit iteration. Example (Change all QNames from prefix abc to xyz and new namespace URI for node $root and its decendents.) for $node in $root//abc:* let $localname := fn:local-name($node), $newqname := fn:concat("xyz:", $localname) return rename node $node as fn:qname(" $newqname), for $attr in $node/@abc:* let $attrlocalname := fn:local-name($attr), $attrnewqname := fn:concat("xyz:", $attrlocalname) return rename node $attr as fn:qname(" $attrnewqname) Marc H. Scholl (DBIS, Uni KN) XML and Databases Winter 7/8 598

13 . Updating XML Trees XQuery Update Facility Node transformation... creates modified copies of existing nodes. Each copied node obtains a new node identity. The resulting XDM instance can contain both, newly created and previously existing nodes. Node transformation is a non-updating expression, since it does not modify existing nodes! Syntax TransformExpr ::= "copy" "$"VarName ":=" ExprSingle ("," "$"VarName ":=" ExprSingle)* "modify" ExprSingle "return" ExprSingle Idea: Bind variables of copy clause (non-updating expressions), update copies (only!) as per modify clause, construct result by return (copied/modified and/or other nodes). Marc H. Scholl (DBIS, Uni KN) XML and Databases Winter 7/8 599

14 . Updating XML Trees XQuery Update Facility Node transformation: Examples Return a sequence consisting of all employee elements that have Java as a skill, excluding their salary child-elements. for $e in //employee[skill = "Java"] return copy $je := $e modify delete node $je/salary return $je Copy a node, modify copy, then return original and modified copy. let $oldx := /a/b/x return copy $newx := $oldx modify (rename node $newx as "newx", replace value of node $newx by $newx * ) return ($oldx, $newx) N.B. Underlying persistent data not changed by these examples! Marc H. Scholl (DBIS, Uni KN) XML and Databases Winter 7/8 6

15 . Updating XML Trees XQuery Update Facility On the semantics of the XQuery Update Facility Formally specifying the exact semantics of the XQuery UF is non-trivial for several reasons: Formal update semantics are always a lot more involved than retrieval semantics. Updates and bulk operations do not go together well (cf. SQL set-oriented updates). XUF uses a notion of snapshots and pending update lists to work around some of the subtleties. The details are beyond the scope of our current section. Marc H. Scholl (DBIS, Uni KN) XML and Databases Winter 7/8 6

16 . Impact on XPath Accelerator Encoding Text node updates Obviously, replacing the value of a text (or attribute, comment, processing instruction) node has little impact on the XML representation. Replacing text by text <a> foo bar </a> replace text bar by foo <a> foo foo </a> Marc H. Scholl (DBIS, Uni KN) XML and Databases Winter 7/8 6

17 . Impact on XPath Accelerator Encoding Text node updates Translated into, e.g., the XPath Accelerator representation, we see that Replacing text nodes by text nodes has local impact only on the pre/post encoding of the updated tree. The update leads to a local relational update pre post text NULL NULL foo NULL bar pre post text NULL NULL foo NULL foo Similar observations can be made for updates on comment and processing instruction nodes. Marc H. Scholl (DBIS, Uni KN) XML and Databases Winter 7/8 6

18 . Impact on XPath Accelerator Encoding Structural updates Inserting a new subtree <a> <c><d/><e/></c> <f><g/> <h><j/></h> </f> </a> insert node <k><l/><m/></k> into /a/f/g <a> <c><d/><e/></c> <f><g><k><l/><m/></k></g> <h><j/></h> </f> </a> Question: What are the effects w.r.t. our structure encoding...? Marc H. Scholl (DBIS, Uni KN) XML and Databases Winter 7/8 6

19 . Impact on XPath Accelerator Encoding Insertion: Global impact on encoding Global shifts in the pre/post Plane post a 5 b c e, f h i j g d 5 pre post post+ a f g 5 b c e, pre+;post+ h k l m i j d 5 pre Marc H. Scholl (DBIS, Uni KN) XML and Databases Winter 7/8 65

20 . Impact on XPath Accelerator Encoding Insertion: Global impact on pre/post plane Insert a subtree of n nodes below parent element v post(v) post(v) + n v v/following::node(): pre(v ) pre(v ) + n; post(v ) post(v ) + n v v/ancestor::node(): post(v ) post(v ) + n Cost (tree of N nodes) O(N) + O(log N) }{{}}{{} Update cost is not so much a problem of cost but of locking. Why? Marc H. Scholl (DBIS, Uni KN) XML and Databases Winter 7/8 66

21 . Impacts on Other Encoding Schemes Updates and fixed-width encodings Theoretical result [Milo et.al., PODS ] There is a sequence of updates (subtree insertions) for any persistent 59 tree encoding scheme E, such that E needs labels of length Ω(N) to encode the resulting tree of N nodes. Fixed-width tree encodings (like XPath Accelerator) are inherently static. Non-solutions: Gaps in the encoding, encodings based on decimal fractions. 59 A node keeps its initial encoding label even if its tree is updated. Marc H. Scholl (DBIS, Uni KN) XML and Databases Winter 7/8 67

22 . Impacts on Other Encoding Schemes ORDPATH Encoding A variable-width tree encoding: ORDPATH Here we look at a particular variant of a hierarchical numbering scheme, optimized for updates. The ORDPATH encoding (used in MS SQL Server TM ) assigns node labels of variable length. ORDPATH labels for an XML fragment The fragment root receives label. The nth (n =,,... ) child of a parent node labelled p receives label p ( n ). ORDPATH looks like the hierarchical numbering scheme we have seen earlier, but it uses only odd numbers (in the first place). Internally, ORDPATH labels are not stored as -separated ordinals but using a prefix-encoding (similarities with Unicode). Marc H. Scholl (DBIS, Uni KN) XML and Databases Winter 7/8 68

23 . Impacts on Other Encoding Schemes ORDPATH Encoding ORDPATH encoding: Example ORDPATH encoding of a sample XML fragment <a> a <c> <d/><e/> b c h <f><g/></f> 5 </c> d 5 e f i j k <h> <j/><k/> </h> 5 g </a> Note: Lexicographic order of ORDPATH labels document order Clustered index on ORDPATH labels will be helpful. Marc H. Scholl (DBIS, Uni KN) XML and Databases Winter 7/8 69

24 . Impacts on Other Encoding Schemes ORDPATH Encoding ORDPATH: Insertion between siblings In ORDPATH, the insertion of new nodes between two existing sibling nodes is referred to as careting in (caret ˆ= insertion mark, ). ORDPATH: node insertion Let (v,..., v n ) denote a sequence of nodes to be inserted between two existing sibling nodes with labels p s and p (s + ), s odd. After insertion, the new label of v i is p (s + ) ( i ). Label p (s + ) is referred to as a caret. N.B. That is: the even numbers only serve as carets! Marc H. Scholl (DBIS, Uni KN) XML and Databases Winter 7/8 6

25 . Impacts on Other Encoding Schemes ORDPATH Encoding ORDPATH: Insertion between siblings (Example) Insertion of (<l/>, <m/>) between <j/> and <k/> a b a b h c 5 d e f d e f i 5 5 j k 5 5 g h c 5 i j 5 k g 5 l m 5 Marc H. Scholl (DBIS, Uni KN) XML and Databases Winter 7/8 6

26 . Impacts on Other Encoding Schemes ORDPATH Encoding ORDPATH: Insertion between siblings ORDPATH: Insertions at arbitrary locations? a b h c 5 d e f i j 5 k g 5 l m 5 Determine ORDPATH label of new node v inserted to the right of <k/>, to the left of , between <j/> and <l/>, between <l/> and <m/>, Marc H. Scholl (DBIS, Uni KN) XML and Databases Winter 7/8 6

27 . Impacts on Other Encoding Schemes ORDPATH Encoding Processing XQuery and ORDPATH Is ORDPATH a suitable encoding E? Mapping core operations of the XQuery processing model to operations on ORDPATH labels: v/parent::node() Let p m n denote v s label (n is odd). If the rightmost ordinal (m) is even, remove it. Goto. In other words: the carets ( ) do not count for ancestry. v/descendant::node() Let p n denote v s label (n is odd). Perform a lexicographic index range scan from p n to p (n + ) the virtual following sibling of v. Marc H. Scholl (DBIS, Uni KN) XML and Databases Winter 7/8 6

28 . Impacts on Other Encoding Schemes ORDPATH Encoding ORDPATH: Variable-length node encoding Using ( byte) integers for all numbers in the hierarchical numbering scheme is an obvious waste of space! Fewer (and variable number of) bits are typically sufficient; they may bear the risk of running out of new numbers, though. In that case, even ORDPATH cannot avoid renumbering. In principle, though, no bounded representation can absolutely avoid the need for renumbering. Several approaches have been proposed so as to alleviate the problem, for instance: use a variable number of bits/bytes, akin to Unicode, apply some (order-preserving) hashing schemes to shorten the numbers,... Marc H. Scholl (DBIS, Uni KN) XML and Databases Winter 7/8 6

29 . Impacts on Other Encoding Schemes ORDPATH Encoding ORDPATH: Variable-length node encoding For a MB XML sample document, the authors of ORDPATH observed label lenghts between 6 and bytes (using Unicode-like compact representations). Since ORDPATH labels encode root-to-node paths, node labels share common prefixes. ORDPATH labels of <l/> and <m/> 5 5 Label comparisons often need to inspect encoding bits at the far right. MS SQL Server TM employs further path encodings organized in reverse (node-to-root) order. Marc H. Scholl (DBIS, Uni KN) XML and Databases Winter 7/8 65

30 . Impacts on Other Encoding Schemes ORDPATH Encoding ORDPATH vs. XPath Accelerator Fixed-length node IDs (such as, e.g., preorder ranks) typically fit into CPU registers, while variable-length IDs do not. Comparison of fixed-length IDs (e.g., to determine document order or within a staircase-join) is a single, fast CPU instruction, while comparing variable-length IDs requires additional overhead (such as, a loop, for instance). Pre-/Post-Order ranks need to be updated after each insertion/deletion, for each node to the right of the update position while hierarchical nodeids are stable w.r.t. those updates. In both cases, one cannot completely eliminate the need for renumbering. Marc H. Scholl (DBIS, Uni KN) XML and Databases Winter 7/8 66

31 . What can Pathfinder/MonetDB do? What can Pathfinder/MonetDB do? Divide tables into (logical) pages, leave room at end of pages, use new pos attribute instead of pre, only append new pages to old ones as needed, use clever memory mapping to reassure desired page order. [XIME-P 5] Marc H. Scholl (DBIS, Uni KN) XML and Databases Winter 7/8 67

32 . What can Pathfinder/MonetDB do? Updates in Pathfinder/MonetDB When the XML tree structure is changed, the pre/post-values need to be adapted. a b c d e f g h i j pre size level a 9 f following 8 h 7 j 6 i 5 ancestor g b preceding c e d descendant pre <xupdate:append select= /a/f/g > <k> <l/> <m/> </k> a 9 f following 8 7 size+ h 6 pre+ i j 5 ancestor g b preceding k m c new e l nodes d descendant pre pre size level a b c d e f g h i j k l m Figure : The impact of Structural Updates on pre/size/level XML Storage We use pre size level and derive post from these.. UPDATEABLE XML ENCODING pping is particularly well-supported in MonetDB, In the MonetDB/XQuery implementation, pre is not explicitly stored s the pre column using the void type. A void Figure shows the changes introduced in Monet lds a densely ( void ascending attribute). node sequence (,,,...). to handle structural updates in the pre/size/level columns are actually not stored at all: they take The key observations are: More importantly, size andcareful level design may and need implemenelational tables in MonetDB allows to support the table is called pos/size/level now. to be adapted, too. voidmarc values H. Scholl using(dbis, positional Uni KN) algorithms, such XMLasand Databases Winter 7/8 68

33 . What can Pathfinder/MonetDB do? Updatable MonetDB/XQuery Representation Read Only pre size level node 9 a b c d e f g h i j pagesize = 8 unused space: level = NULL size set to unite consecutive space vs. Updatable Representation pos size level a b c d e f g <xupdate:append select= /a/f/g > <k> <l/> <m/> </k> first try to handle the insert inside a page h i j if full, append pages (NULL padded) pos size level pre/size/level is a memory mapped view with pages in logical order pre is a virtual column (void), therefore it adapts automatically pre size level a b c d e 7 f g k l m 5 h i j text, Each ta white colu use fast p schema ov In the placed by limits nod tuples are table, whi structural attribute a unique time. It is shredding (the latte The fin rized in F maintain this pageo mapping Marc H. Scholl (DBIS, Uni KN) XML and Databases Winter 7/8 69

Updating XML documents

Grindei Manuela Lidia Updating XML documents XQuery Update XQuery is the a powerful functional language, which enables accessing different nodes of an XML document. However, updating could not be done