Gestão e Tratamento de Informação

Transcription

1 Departamento de Engenharia Informática 2013/2014 Gestão e Tratamento de Informação 1st Project Deadline at 25 Oct :: Online submission at IST/Fénix The SIGMOD Record 1 journal is a quarterly publication of the Special Interest Group on Management of Data of the Association for Computing Machinery (i.e., the ACM SIGMOD 2, which includes researchers that work on the development and application of database technology, supporting a diverse range of data management needs). The XML document available from the Uniform Resource Locator (URL) corresponding to contains a sample from all the articles published in ACM SIGMOD Record. Figure 1 illustrates the contents of the XML document that is available from the above URL. <?xml version='1.0' encoding='utf-8'?> <sigmod> <articles> <article> <title>message Forms in Computer Networks.</title> <initpage>38</initpage> <endpage>42</endpage> <volume>1</volume> <issue>1</issue> <authors> <authorcode>8</authorcode> <authorcode>9</authorcode> </authors> </article> <!-- list of remaining articles --> </articles> <authors> <author code="1">donald J. Hatfield</author> <author code="3">g. A. Gibson</author> <!-- list of remaining authors --> </authors> </sigmod> Figure 1 : Sample from the XML document used in the exercises. Exercise Create an XML Schema for validating the XML document with the sample of articles published on SIGMOD Record. When developing the XML Schema, take the following particular aspects into consideration: Ensure that the developed XML Schema uses global data types, instead of local type definitions encapsulated within the elements, in order to promote modularity and code re- usage. Ensure that the data types for the contents of elements named volume correspond to positive integer numbers record/ IST/DEI Page 1 of 8

2 Ensure that author names contain at least one uppercase letter. Ensure that the attribute named code is of mandatory occurrence in the elements named author. Assume that the elements descending from article can appear with any order Extend the XML Schema from the previous exercise in order to take referential integrity into account, when validating the document. Specifically, verify if the authorcode elements, used in the definition of articles, refer to authors that exist within the XML document. Exercise 2 Create an XML Transformation (i.e., an XSLT document) for representing the XML sample of articles from SIGMOD Record as an XHTML document that can be displayed in a Web browser. The developed XSLT should use structural recursion to process the input XML document, including at least two different templates. The resulting XHTML document should contain one separate table for each volume/issue of the journal, describing the corresponding articles. Each of these XHTML tables should be preceded by a string of the form Papers from issue XX of volume XX, encoding the corresponding volume/issue of the journal, as a description for the table. Each XHTML table should show information related to the titles of papers, their first authors, and the range of pages associated to each article, as three separate columns. The titles for the articles should be presented in uppercase, and the range of pages should be presented as a string of the form [start- page ; end- page] (i.e., present the start and end pages between rectangular brackets, separated by a semi- comma). The name for the author of each article should be presented as a string with the author name, for the case of articles with a single author, or as string of the form Name for First Author et al., for the case of articles with more than one author (i.e., use the suffix et al. in these cases). Exercise 3 Create XPath expressions for addressing each of the following information needs. 1. Find how many articles were published in the first issue of volume 3 for SIGMOD Record. 2. Find the papers related to relational databases (i.e., the papers whose titles contain the word relational together with some variation of the word data). 3. Find the names of all co- authors of E. F. Codd (i.e. the names of all authors who have published at least one article together with the author named E. F. Codd). Only the textual contents corresponding to the author names should be presented, and the answers should use only uppercase letters. 4. Find the title(s) of the paper(s) with the highest number of co- authors. IST/DEI Page 2 of 8

3 Exercise 4 Create XQuery FLWOR expressions for addressing each of the following information needs. Although XPath expressions can be used as part of each answer, note that in this exercise you should always write the queries using XQuery FLWOR expressions, eventually also using XQuery updating expressions whenever the exercise involves updating the XML dataset. 1. Find the name(s) of the author(s) who have written more than one article together with the author named Isabel F. Cruz. 2. Find the top five articles with more pages of content, showing the titles and authors of each of these papers according to the format exemplified bellow. When producing the list, you may ignore all those articles that do not contain information regarding the starting and ending page, and the articles where these elements do not contain an integer value. <papers> <paper title="title for paper 1" authors="author-name-1 and AUTHOR-NAME-2" num-pages="5" /> <paper title="title for paper 1" authors="author-name-1 and AUTHOR-NAME-2" num-pages="4" /> <!-- remaining papers in the list of top five papers --> </papers> Figure 2 : Example of the format produced as output by the XQuery expression. 3. Change the XML dataset in order to explicitly store (i) an attribute associated to the volume elements, encoding the number of issues published in each volume, and (ii) an attribute associated to the issue elements, encoding the number of articles published in each issue number of each volume. 4. Change the XML dataset so as to remove all papers with less than two authors, removing also all authors that have not published any of the papers in the resulting data set. Exercise 5 Consider the following two tree structures (see Figure 3), each representing an XML element encoding information about papers published on ACM SIGMOD Record. Figure 3 : Two example tree structures. IST/DEI Page 3 of 8

4 Compute the similarity (i.e., the number of matching nodes) and the alignment between both trees, using the Simple Tree Matching algorithm, and considering that two nodes can be aligned if they share the same name/content. Present all the calculations involved in finding the number of matching nodes in the trees. Present also the alignment between the trees, with basis on the results from the Simple Tree Matching algorithm. Submitting the project The solutions to the project should be submitted through the Fénix system, in the form of a.zip file containing a PDF report with the solutions for the exercises, as well as individual text files with the solutions for each of the exercises (i.e., documents with the XML, XSD, XSLT or XPath/XQuery code). In the course Webpage, you can find a Microsoft word template for the project report. In the theoretical class following the electronic submission, a printed copy of the report, with the solutions for each exercise, should also be delivered. We will not accept deliveries through e- mail, with reports not conforming to the supplied template, or without the text files with the solutions for the exercises. Good luck! IST/DEI Page 4 of 8

5 Solutions for Exercise 1.1 and 1.2 The schema for validating the XML document is shown bellow. The solution to Exercise 1.2 corresponds to the key and keyref definitions that are included in the definition of the XML element named sigmod. <xsd:schema xmlns:xsd=" <xsd:element name="sigmod"> <xsd:complextype> <xsd:sequence> <xsd:element name="articles"> <xsd:complextype> <xsd:sequence> <xsd:element name="article" type="articletype" minoccurs="1" maxoccurs="unbounded"/> </xsd:sequence> </xsd:element> <xsd:element name="authors" type="authortype" minoccurs="1" maxoccurs="unbounded"> </xsd:element> </xsd:sequence> <xsd:key name="mykey"> <xsd:selector xpath="authors/author" /> <xsd:field /> </xsd:key> <xsd:keyref name="mykeyref" refer="mykey"> <xsd:selector xpath="article/authors" /> <xsd:field xpath="authorcode" /> </xsd:keyref> </xsd:element> <xsd:complextype name="articletype"> <xsd:all> <xsd:element name="title" type="xsd:string" /> <xsd:element name="initpage" type="xsd:string" /> <xsd:element name="endpage" type="xsd:string" /> <xsd:element name="volume" type="xsd:positiveinteger" /> <xsd:element name="issue" type="xsd:string" /> <xsd:element name="authors" type="authortype2"/> </xsd:all> <xsd:simpletype name="uppercasename"> <xsd:restriction base="xsd:string"><xsd:pattern value=".*[a-z].*"/></xsd:restriction> </xsd:simpletype> <xsd:complextype name="authortype"> <xsd:sequence> <xsd:element name="author" minoccurs="0" maxoccurs="unbounded"> <xsd:complextype> <xsd:simplecontent> <xsd:extension base="uppercasename"> <xsd:attribute name="code" type="xsd:integer" use="required" /> </xsd:extension> </xsd:simplecontent> </xsd:element> </xsd:sequence> <xsd:complextype name="authortype2"> <xsd:sequence> <xsd:element name="authorcode" minoccurs="0" maxoccurs="unbounded" type="xsd:string" /> </xsd:sequence> </xsd:schema> IST/DEI Page 5 of 8

6 Solution for Exercise 2 <xsl:stylesheet version="1.0" xmlns:xsl=" <xsl:template match="/"> <html> <head><title>papers published on SIGMOD Record</title></head> <body><ul><xsl:apply-templates /></ul></body> </html> <xsl:template match="sigmod"> <xsl:for-each select="//article[not(volume = following::article/volume)]/volume/text()"> <xsl:variable name="vol" select="." /> <xsl:for-each select="//article[volume=$vol][not(issue = following::article[volume=$vol]/issue)]/issue/text()"> <xsl:variable name="issue" select="." /> <li> Papers from issue <xsl:value-of select="$issue" /> of volume <xsl:value-of select="$vol" /> <br/> <table border="1"> <tr><th>title</th><th>author</th><th>pages</th></tr> <xsl:for-each select="//article[volume=$vol][issue=$issue]"> <tr><xsl:apply-templates /></tr> </xsl:for-each> </table> </li> </xsl:for-each> </xsl:for-each> <xsl:template match="title"> <xsl:variable name="smallcase" select="'abcdefghijklmnopqrstuvwxyz'" /> <xsl:variable name="uppercase" select="'abcdefghijklmnopqrstuvwxyz'" /> <td><xsl:value-of select="translate(text(), $smallcase, $uppercase)" /></td> <xsl:template match="authors"> <td> <xsl:variable name="code" select="./authorcode[1]/text()" /> <xsl:choose> <xsl:when test="count(.//authorcode) = 1"> <xsl:value-of select="//author[@code=$code]/text()" /> </xsl:when> <xsl:when test="count(.//authorcode) > 1"> <xsl:value-of select="//author[@code=$code]/text()" /> et al. </xsl:when> <xsl:otherwise></xsl:otherwise> </xsl:choose> </td> <xsl:template match="endpage"> <td><xsl:value-of select="concat('[',../initpage/text(),' ; ', text(), ']')" /></td> <xsl:template match="text()"> </xsl:stylesheet> Solutions for Exercise 3 The four XPath expressions shown below correspond to the solution of each of the proposed exercises. count( doc("sigmod-data.xml")//article[volume=3 and issue=1] ) doc("sigmoddata.xml")/sigmod/articles/article[matches(title/text(),"[rr]elational")][matches(title/text(),"[dd]a ta")] doc("sigmod-data.xml")//author[./text()!="e. F. Codd"] [@code=//article[.//authorcode = //author[text()="e. F. Codd"]/@code] /authors/authorcode]/upper-case(text()) doc("sigmod-data.xml")//article[count(.//authorcode) = max(doc("sigmod-data.xml")//article/count(.//authorcode))]/title IST/DEI Page 6 of 8

7 Solution for Exercise 4.1 let $doc := doc("sigmod-data.xml") let $code := $doc/sigmod/authors/author[text()="isabel F. let $articles := $doc//article[./authors/authorcode = $code] let $coauthors := for $a in $articles//authorcode where count($articles[.//authorcode = $a]) > 1 return $a/text() return $doc//author[@code!= $code Solution for Exercise 4.2 <articles> { let $articles := let $doc := doc("sigmod-data.xml") for $a in $doc//article[matches(endpage/text(),"^[0-9]+$")] [matches(initpage/text(),"^[0-9]+$")] let $pages := $a/endpage - $a/initpage + 1 let $authors := string-join($doc//author[@code = $a//authorcode/text()]/upper-case(.), " and ") order by $pages descending return <article title="{$a/title}" authors="{$authors}" numpages="{$pages}" /> return $articles[position() < 6] } </articles> Solution for Exercise 4.3 copy $doc := doc("sigmod-data.xml") modify ( for $a in $doc//volume let $num-issues := count($doc//article[volume=$a]//issue) return insert node attribute num-issues {$num-issues} into $a, for $a in distinct-values($doc//volume), $b in $doc//article[volume=$a]/issue let $num-articles := count($doc//article[volume=$a][issue=$b]) return insert node attribute num-articles {$num-articles} into $b ) return $doc Solution for Exercise 4.4 let $newdoc := copy $doc := doc("sigmod-data.xml") modify delete node $doc//article[count(.//authorcode) < 2] return $doc return copy $doc := doc("sigmod-data.xml") modify ( delete node $doc//article[count(.//authorcode) < 2], for $a in $doc//author where count($newdoc//article[.//authorcode=$a/@code]) < 1 return delete node $a ) return $doc Solution for Exercise 5 The number of matching nodes between both trees is 8. Descendants from paper title year journal authors publisher pages 0 title journal year authors pages IST/DEI Page 7 of 8

8 Descendants from title XML XML 0 1 Descendants from year Descendants from journal volume issue title vol/issue title 0 2 Descendants from title TXML TXML 0 1 Descendants from authors inácio et al. inácio rodrigo santos Descendants from pages s The aligned nodes between both trees are as follows: A:paper B:paper A:title B:title A:XML B:XML A:journal B:journal A:title B:title A:TXML B:TXML A:authors B:authors A:pages B:pages IST/DEI Page 8 of 8