Skip to content

X
XPath Parser
Commit 84170241 authored by David BAELDE's avatar David BAELDE
Browse files
Options

Factorize list of XPath std funs

parent 1567569e
Hide whitespace changes
Inline Side-by-side
Showing with 239 additions and 456 deletions
tools/extensions/XPathStd.pm 0 → 100644
View file @ 84170241
package XPathStd;
use Exporter;
our @ISA= qw( Exporter );
our @EXPORT_OK = qw( functions );
our @EXPORT = qw( functions );
@functions = (
"abs",
"acos",
"add-dayTimeDurations",
"add-dayTimeDuration-to-date",
"add-dayTimeDuration-to-dateTime",
"add-dayTimeDuration-to-time",
"add-yearMonthDurations",
"add-yearMonthDuration-to-date",
"add-yearMonthDuration-to-dateTime",
"adjust-dateTime-to-timezone",
"adjust-date-to-timezone",
"adjust-time-to-timezone",
"analyze-string",
"asin",
"atan",
"atan2",
"available-environment-variables",
"avg",
"base64Binary-equal",
"base-uri",
"boolean",
"boolean-equal",
"boolean-greater-than",
"boolean-less-than",
"ceiling",
"codepoint-equal",
"codepoints-to-string",
"collection",
"compare",
"concat",
"concatenate",
"contains",
"cos",
"count",
"current-date",
"current-dateTime",
"current-time",
"data",
"date-equal",
"date-greater-than",
"date-less-than",
"dateTime",
"dateTime-equal",
"dateTime-greater-than",
"dateTime-less-than",
"day-from-date",
"day-from-dateTime",
"days-from-duration",
"dayTimeDuration-greater-than",
"dayTimeDuration-less-than",
"deep-equal",
"default-collation",
"distinct-values",
"divide-dayTimeDuration",
"divide-dayTimeDuration-by-dayTimeDuration",
"divide-yearMonthDuration",
"divide-yearMonthDuration-by-yearMonthDuration",
"doc",
"doc-available",
"document-uri",
"duration-equal",
"element-with-id",
"empty",
"encode-for-uri",
"ends-with",
"environment-variable",
"error",
"escape-html-uri",
"exactly-one",
"except",
"exists",
"exp",
"exp10",
"false",
"filter",
"floor",
"fold-left",
"fold-right",
"for-each",
"for-each-pair",
"format-date",
"format-dateTime",
"format-integer",
"format-number",
"format-time",
"function-arity",
"function-lookup",
"function-name",
"gDay-equal",
"generate-id",
"gMonthDay-equal",
"gMonth-equal",
"gYear-equal",
"gYearMonth-equal",
"has-children",
"head",
"hexBinary-equal",
"hours-from-dateTime",
"hours-from-duration",
"hours-from-time",
"id",
"idref",
"implicit-timezone",
"index-of",
"innermost",
"in-scope-prefixes",
"insert-before",
"intersect",
"iri-to-uri",
"is-same-node",
"lang",
"last",
"local-name",
"local-name-from-QName",
"log",
"log10",
"lower-case",
"matches",
"max",
"min",
"minutes-from-dateTime",
"minutes-from-duration",
"minutes-from-time",
"month-from-date",
"month-from-dateTime",
"months-from-duration",
"multiply-dayTimeDuration",
"multiply-yearMonthDuration",
"name",
"namespace-uri",
"namespace-uri-for-prefix",
"namespace-uri-from-QName",
"nilled",
"node-after",
"node-before",
"node-name",
"normalize-space",
"normalize-unicode",
"not",
"NOTATION-equal",
"number",
"numeric-add",
"numeric-divide",
"numeric-equal",
"numeric-greater-than",
"numeric-integer-divide",
"numeric-less-than",
"numeric-mod",
"numeric-multiply",
"numeric-subtract",
"numeric-unary-minus",
"numeric-unary-plus",
"one-or-more",
"outermost",
"parse-xml",
"parse-xml-fragment",
"path",
"pi",
"position",
"pow",
"prefix-from-QName",
"QName",
"QName-equal",
"remove",
"replace",
"resolve-QName",
"resolve-uri",
"reverse",
"root",
"round",
"round-half-to-even",
"seconds-from-dateTime",
"seconds-from-duration",
"seconds-from-time",
"serialize",
"sin",
"sqrt",
"starts-with",
"static-base-uri",
"string",
"string-join",
"string-length",
"string-to-codepoints",
"subsequence",
"substring",
"substring-after",
"substring-before",
"subtract-dates",
"subtract-dateTimes",
"subtract-dayTimeDuration-from-date",
"subtract-dayTimeDuration-from-dateTime",
"subtract-dayTimeDuration-from-time",
"subtract-dayTimeDurations",
"subtract-times",
"subtract-yearMonthDuration-from-date",
"subtract-yearMonthDuration-from-dateTime",
"subtract-yearMonthDurations",
"sum",
"tail",
"tan",
"time-equal",
"time-greater-than",
"time-less-than",
"timezone-from-date",
"timezone-from-dateTime",
"timezone-from-time",
"to",
"tokenize",
"trace",
"translate",
"true",
"union",
"unordered",
"unparsed-text",
"unparsed-text-available",
"unparsed-text-lines",
"upper-case",
"uri-collection",
"year-from-date",
"year-from-dateTime",
"yearMonthDuration-greater-than",
"yearMonthDuration-less-than",
"years-from-duration",
"zero-or-one"
);
1;
tools/extensions/countfuns.pl
View file @ 84170241
#!/usr/bin/perl #!/usr/bin/perl
# Standard XPATH functions require "./XPathStd.pm";
@stdfuns = (
"abs",
"acos",
"add-dayTimeDurations",
"add-dayTimeDuration-to-date",
"add-dayTimeDuration-to-dateTime",
"add-dayTimeDuration-to-time",
"add-yearMonthDurations",
"add-yearMonthDuration-to-date",
"add-yearMonthDuration-to-dateTime",
"adjust-dateTime-to-timezone",
"adjust-date-to-timezone",
"adjust-time-to-timezone",
"analyze-string",
"asin",
"atan",
"atan2",
"available-environment-variables",
"avg",
"base64Binary-equal",
"base-uri",
"boolean",
"boolean-equal",
"boolean-greater-than",
"boolean-less-than",
"ceiling",
"codepoint-equal",
"codepoints-to-string",
"collection",
"compare",
"concat",
"concatenate",
"contains",
"cos",
"count",
"current-date",
"current-dateTime",
"current-time",
"data",
"date-equal",
"date-greater-than",
"date-less-than",
"dateTime",
"dateTime-equal",
"dateTime-greater-than",
"dateTime-less-than",
"day-from-date",
"day-from-dateTime",
"days-from-duration",
"dayTimeDuration-greater-than",
"dayTimeDuration-less-than",
"deep-equal",
"default-collation",
"distinct-values",
"divide-dayTimeDuration",
"divide-dayTimeDuration-by-dayTimeDuration",
"divide-yearMonthDuration",
"divide-yearMonthDuration-by-yearMonthDuration",
"doc",
"doc-available",
"document-uri",
"duration-equal",
"element-with-id",
"empty",
"encode-for-uri",
"ends-with",
"environment-variable",
"error",
"escape-html-uri",
"exactly-one",
"except",
"exists",
"exp",
"exp10",
"false",
"filter",
"floor",
"fold-left",
"fold-right",
"for-each",
"for-each-pair",
"format-date",
"format-dateTime",
"format-integer",
"format-number",
"format-time",
"function-arity",
"function-lookup",
"function-name",
"gDay-equal",
"generate-id",
"gMonthDay-equal",
"gMonth-equal",
"gYear-equal",
"gYearMonth-equal",
"has-children",
"head",
"hexBinary-equal",
"hours-from-dateTime",
"hours-from-duration",
"hours-from-time",
"id",
"idref",
"implicit-timezone",
"index-of",
"innermost",
"in-scope-prefixes",
"insert-before",
"intersect",
"iri-to-uri",
"is-same-node",
"lang",
"last",
"local-name",
"local-name-from-QName",
"log",
"log10",
"lower-case",
"matches",
"max",
"min",
"minutes-from-dateTime",
"minutes-from-duration",
"minutes-from-time",
"month-from-date",
"month-from-dateTime",
"months-from-duration",
"multiply-dayTimeDuration",
"multiply-yearMonthDuration",
"name",
"namespace-uri",
"namespace-uri-for-prefix",
"namespace-uri-from-QName",
"nilled",
"node-after",
"node-before",
"node-name",
"normalize-space",
"normalize-unicode",
"not",
"NOTATION-equal",
"number",
"numeric-add",
"numeric-divide",
"numeric-equal",
"numeric-greater-than",
"numeric-integer-divide",
"numeric-less-than",
"numeric-mod",
"numeric-multiply",
"numeric-subtract",
"numeric-unary-minus",
"numeric-unary-plus",
"one-or-more",
"outermost",
"parse-xml",
"parse-xml-fragment",
"path",
"pi",
"position",
"pow",
"prefix-from-QName",
"QName",
"QName-equal",
"remove",
"replace",
"resolve-QName",
"resolve-uri",
"reverse",
"root",
"round",
"round-half-to-even",
"seconds-from-dateTime",
"seconds-from-duration",
"seconds-from-time",
"serialize",
"sin",
"sqrt",
"starts-with",
"static-base-uri",
"string",
"string-join",
"string-length",
"string-to-codepoints",
"subsequence",
"substring",
"substring-after",
"substring-before",
"subtract-dates",
"subtract-dateTimes",
"subtract-dayTimeDuration-from-date",
"subtract-dayTimeDuration-from-dateTime",
"subtract-dayTimeDuration-from-time",
"subtract-dayTimeDurations",
"subtract-times",
"subtract-yearMonthDuration-from-date",
"subtract-yearMonthDuration-from-dateTime",
"subtract-yearMonthDurations",
"sum",
"tail",
"tan",
"time-equal",
"time-greater-than",
"time-less-than",
"timezone-from-date",
"timezone-from-dateTime",
"timezone-from-time",
"to",
"tokenize",
"trace",
"translate",
"true",
"union",
"unordered",
"unparsed-text",
"unparsed-text-available",
"unparsed-text-lines",
"upper-case",
"uri-collection",
"year-from-date",
"year-from-dateTime",
"yearMonthDuration-greater-than",
"yearMonthDuration-less-than",
"years-from-duration",
"zero-or-one"
);
unless (@ARGV) { unless (@ARGV) {
print <<HELP; print <<HELP;
...@@ -305,7 +79,7 @@ for $fname (sort { $table{$b} <=> $table{$a} } keys %table) { ...@@ -305,7 +79,7 @@ for $fname (sort { $table{$b} <=> $table{$a} } keys %table) {
"* $n functions (out of $nbfuns) needed to cover 70% of occurrences.\n"; "* $n functions (out of $nbfuns) needed to cover 70% of occurrences.\n";
} }
printf DAT "$n $table{$fname} %.2f\n", 100*$sofar/$total; printf DAT "$n $table{$fname} %.2f\n", 100*$sofar/$total;
if (grep { $fname eq $_ } @stdfuns) { if (grep { $fname eq $_ } @XPathStd::functions) {
$sofar_std+=$table{$fname}; $sofar_std+=$table{$fname};
printf STD "$n $table{$fname} %.2f\n",100*$sofar_std/$total; printf STD "$n $table{$fname} %.2f\n",100*$sofar_std/$total;
} else { } else {
......
tools/extensions/nonstandard.pl
View file @ 84170241
#!/usr/bin/perl #!/usr/bin/perl
require "./XPathStd.pm";
die "Usage: $0 <XML files>\n" unless @ARGV; die "Usage: $0 <XML files>\n" unless @ARGV;
@files = @ARGV; @files = @ARGV;
...@@ -209,235 +211,8 @@ foreach (@unsupported) { ...@@ -209,235 +211,8 @@ foreach (@unsupported) {
$query = "$query)"; $query = "$query)";
$nonsupportedquery = "ast//xqx:functionName[$query]"; $nonsupportedquery = "ast//xqx:functionName[$query]";
@functions = (
"abs",
"acos",
"add-dayTimeDurations",
"add-dayTimeDuration-to-date",
"add-dayTimeDuration-to-dateTime",
"add-dayTimeDuration-to-time",
"add-yearMonthDurations",
"add-yearMonthDuration-to-date",
"add-yearMonthDuration-to-dateTime",
"adjust-dateTime-to-timezone",
"adjust-date-to-timezone",
"adjust-time-to-timezone",
"analyze-string",
"asin",
"atan",
"atan2",
"available-environment-variables",
"avg",
"base64Binary-equal",
"base-uri",
"boolean",
"boolean-equal",
"boolean-greater-than",
"boolean-less-than",
"ceiling",
"codepoint-equal",
"codepoints-to-string",
"collection",
"compare",
"concat",
"concatenate",
"contains",
"cos",
"count",
"current-date",
"current-dateTime",
"current-time",
"data",
"date-equal",
"date-greater-than",
"date-less-than",
"dateTime",
"dateTime-equal",
"dateTime-greater-than",
"dateTime-less-than",
"day-from-date",
"day-from-dateTime",
"days-from-duration",
"dayTimeDuration-greater-than",
"dayTimeDuration-less-than",
"deep-equal",
"default-collation",
"distinct-values",
"divide-dayTimeDuration",
"divide-dayTimeDuration-by-dayTimeDuration",
"divide-yearMonthDuration",
"divide-yearMonthDuration-by-yearMonthDuration",
"doc",
"doc-available",
"document-uri",
"duration-equal",
"element-with-id",
"empty",
"encode-for-uri",
"ends-with",
"environment-variable",
"error",
"escape-html-uri",
"exactly-one",
"except",
"exists",
"exp",
"exp10",
"false",
"filter",
"floor",
"fold-left",
"fold-right",
"for-each",
"for-each-pair",
"format-date",
"format-dateTime",
"format-integer",
"format-number",
"format-time",
"function-arity",
"function-lookup",
"function-name",
"gDay-equal",
"generate-id",
"gMonthDay-equal",
"gMonth-equal",
"gYear-equal",
"gYearMonth-equal",
"has-children",
"head",
"hexBinary-equal",
"hours-from-dateTime",
"hours-from-duration",
"hours-from-time",
"id",
"idref",
"implicit-timezone",
"index-of",
"innermost",
"in-scope-prefixes",
"insert-before",
"intersect",
"iri-to-uri",
"is-same-node",
"lang",
"last",
"local-name",
"local-name-from-QName",
"log",
"log10",
"lower-case",
"matches",
"max",
"min",
"minutes-from-dateTime",
"minutes-from-duration",
"minutes-from-time",
"month-from-date",
"month-from-dateTime",
"months-from-duration",
"multiply-dayTimeDuration",
"multiply-yearMonthDuration",
"name",
"namespace-uri",
"namespace-uri-for-prefix",
"namespace-uri-from-QName",
"nilled",
"node-after",
"node-before",
"node-name",
"normalize-space",
"normalize-unicode",
"not",
"NOTATION-equal",
"number",
"numeric-add",
"numeric-divide",
"numeric-equal",
"numeric-greater-than",
"numeric-integer-divide",
"numeric-less-than",
"numeric-mod",
"numeric-multiply",
"numeric-subtract",
"numeric-unary-minus",
"numeric-unary-plus",
"one-or-more",
"outermost",
"parse-xml",
"parse-xml-fragment",
"path",
"pi",
"position",
"pow",
"prefix-from-QName",
"QName",
"QName-equal",
"remove",
"replace",
"resolve-QName",
"resolve-uri",
"reverse",
"root",
"round",
"round-half-to-even",
"seconds-from-dateTime",
"seconds-from-duration",
"seconds-from-time",
"serialize",
"sin",
"sqrt",
"starts-with",
"static-base-uri",
"string",
"string-join",
"string-length",
"string-to-codepoints",
"subsequence",
"substring",
"substring-after",
"substring-before",
"subtract-dates",
"subtract-dateTimes",
"subtract-dayTimeDuration-from-date",
"subtract-dayTimeDuration-from-dateTime",
"subtract-dayTimeDuration-from-time",
"subtract-dayTimeDurations",
"subtract-times",
"subtract-yearMonthDuration-from-date",
"subtract-yearMonthDuration-from-dateTime",
"subtract-yearMonthDurations",
"sum",
"tail",
"tan",
"time-equal",
"time-greater-than",
"time-less-than",
"timezone-from-date",
"timezone-from-dateTime",
"timezone-from-time",
"to",
"tokenize",
"trace",
"translate",
"true",
"union",
"unordered",
"unparsed-text",
"unparsed-text-available",
"unparsed-text-lines",
"upper-case",
"uri-collection",
"year-from-date",
"year-from-dateTime",
"yearMonthDuration-greater-than",
"yearMonthDuration-less-than",
"years-from-duration",
"zero-or-one"
);
$query = '(@xqx:prefix != \'fn\') or (true()'; $query = '(@xqx:prefix != \'fn\') or (true()';
foreach (@functions) { foreach (@XPathStd::functions) {
$query = "$query and text() != '$_'"; $query = "$query and text() != '$_'";
} }
$query = "$query)"; $query = "$query)";
......
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment