Changeset: 1acce398fc2c for MonetDB
URL: http://dev.monetdb.org/hg/MonetDB?cmd=changeset;node=1acce398fc2c
Modified Files:
monetdb5/extras/jaql/Tests/All
monetdb5/extras/jaql/Tests/join00.mal
monetdb5/extras/jaql/Tests/join00.stable.err
monetdb5/extras/jaql/Tests/join00.stable.out
monetdb5/extras/jaql/Tests/join01.mal
monetdb5/extras/jaql/jaql.c
monetdb5/extras/jaql/jaql.h
monetdb5/extras/jaql/jaqlgencode.c
Branch: jacqueline
Log Message:
jaql: implemented join operation
currently var.* notation cannot be used
preserve is left as an exercise for the future as well
diffs (truncated from 1133 to 300 lines):
diff --git a/monetdb5/extras/jaql/Tests/All b/monetdb5/extras/jaql/Tests/All
--- a/monetdb5/extras/jaql/Tests/All
+++ b/monetdb5/extras/jaql/Tests/All
@@ -6,6 +6,7 @@ json03
expand00
filter00
transform00
+join00
sort00
top00
variable00
diff --git a/monetdb5/extras/jaql/Tests/join00.mal
b/monetdb5/extras/jaql/Tests/join00.mal
new file mode 100644
--- /dev/null
+++ b/monetdb5/extras/jaql/Tests/join00.mal
@@ -0,0 +1,8 @@
+jaql.x("users = [ {\"name\": \"Jon Doe\", \"password\": \"asdf1234\", \"id\":
1}, {\"name\": \"Jane Doe\", \"password\": \"qwertyui\", \"id\": 2}, {\"name\":
\"Max Mustermann\", \"password\": \"q1w2e3r4\", \"id\": 3} ];");
+jaql.x("pages = [ {\"userid\": 1, \"url\":\"code.google.com/p/jaql/\"},
{\"userid\": 2, \"url\":\"www.cnn.com\"}, {\"userid\": 1,
\"url\":\"java.sun.com/javase/6/docs/api/\"} ];");
+
+jaql.x("join users, pages where users.id == pages.userid into {users.name,
pages.url};");
+jaql.x("join u in users, p in pages where u.id == p.userid into {u.name,
p.url};");
+# TODO: preserve
+#jaql.x("join preserve u in users, p in pages where u.id == p.userid into
{u.name, p.url};");
+
diff --git a/monetdb5/extras/jaql/Tests/join00.stable.err
b/monetdb5/extras/jaql/Tests/join00.stable.err
new file mode 100644
--- /dev/null
+++ b/monetdb5/extras/jaql/Tests/join00.stable.err
@@ -0,0 +1,31 @@
+stderr of test 'join00` in directory 'extras/jaql` itself:
+
+
+# 17:33:31 >
+# 17:33:31 > "mserver5" "--debug=10" "--set" "gdk_nr_threads=0" "--set"
"gdk_dbfarm=/net/sofia.ins.cwi.nl/export/scratch1/fabian/tmp/mtest-jacqueline-sofia.ins.cwi.nl/five/dbfarm"
"--set" "mapi_open=true" "--set" "mapi_port=34627" "--set" "monet_prompt="
"--trace" "--forcemito" "--set" "mal_listing=2" "--dbname=mTests_extras_jaql"
"join00.mal"
+# 17:33:31 >
+
+# builtin opt gdk_dbname = demo
+# builtin opt gdk_dbfarm =
/ufs/fabian/scratch/ssd/monetdb/jacqueline/program-x86_64/var/lib/monetdb5/dbfarm
+# builtin opt gdk_debug = 0
+# builtin opt gdk_alloc_map = no
+# builtin opt gdk_vmtrim = yes
+# builtin opt monet_prompt = >
+# builtin opt monet_daemon = no
+# builtin opt mapi_port = 50000
+# builtin opt mapi_open = false
+# builtin opt mapi_autosense = false
+# builtin opt sql_optimizer = default_pipe
+# builtin opt sql_debug = 0
+# cmdline opt gdk_nr_threads = 0
+# cmdline opt gdk_dbfarm =
/net/sofia.ins.cwi.nl/export/scratch1/fabian/tmp/mtest-jacqueline-sofia.ins.cwi.nl/five/dbfarm
+# cmdline opt mapi_open = true
+# cmdline opt mapi_port = 34627
+# cmdline opt monet_prompt =
+# cmdline opt mal_listing = 2
+# cmdline opt gdk_dbname = mTests_extras_jaql
+
+# 17:33:31 >
+# 17:33:31 > "Done."
+# 17:33:31 >
+
diff --git a/monetdb5/extras/jaql/Tests/join00.stable.out
b/monetdb5/extras/jaql/Tests/join00.stable.out
new file mode 100644
--- /dev/null
+++ b/monetdb5/extras/jaql/Tests/join00.stable.out
@@ -0,0 +1,33 @@
+stdout of test 'join00` in directory 'extras/jaql` itself:
+
+
+# 17:33:31 >
+# 17:33:31 > "mserver5" "--debug=10" "--set" "gdk_nr_threads=0" "--set"
"gdk_dbfarm=/net/sofia.ins.cwi.nl/export/scratch1/fabian/tmp/mtest-jacqueline-sofia.ins.cwi.nl/five/dbfarm"
"--set" "mapi_open=true" "--set" "mapi_port=34627" "--set" "monet_prompt="
"--trace" "--forcemito" "--set" "mal_listing=2" "--dbname=mTests_extras_jaql"
"join00.mal"
+# 17:33:31 >
+
+# MonetDB 5 server v11.8.0 "jacqueline-0c67cb41e11b+"
+# Serving database 'mTests_extras_jaql', using 8 threads
+# Compiled for x86_64-pc-linux-gnu/64bit with 64bit OIDs dynamically linked
+# Found 15.662 GiB available main-memory.
+# Copyright (c) 1993-July 2008 CWI.
+# Copyright (c) August 2008-2012 MonetDB B.V., all rights reserved
+# Visit http://www.monetdb.org/ for further information
+# Listening for connection requests on mapi:monetdb://sofia.ins.cwi.nl:34627/
+# MonetDB/GIS module loaded
+# MonetDB/SQL module loaded
+# MonetDB/DataCell module not loaded: MALException:jaql.context:JAQL
environment not found
+function user.main():void;
+ jaql.x("users = [ {\"name\": \"Jon Doe\", \"password\": \"asdf1234\",
\"id\": 1}, {\"name\": \"Jane Doe\", \"password\": \"qwertyui\", \"id\": 2},
{\"name\": \"Max Mustermann\", \"password\": \"q1w2e3r4\", \"id\": 3} ];");
+ jaql.x("pages = [ {\"userid\": 1, \"url\":\"code.google.com/p/jaql/\"},
{\"userid\": 2, \"url\":\"www.cnn.com\"}, {\"userid\": 1,
\"url\":\"java.sun.com/javase/6/docs/api/\"} ];");
+ jaql.x("join users, pages where users.id == pages.userid into {users.name,
pages.url};");
+ jaql.x("join u in users, p in pages where u.id == p.userid into {u.name,
p.url};");
+# TODO: preserve
+#jaql.x("join preserve u in users, p in pages where u.id == p.userid into
{u.name, p.url};");
+end main;
+[ { "name": "Jon Doe", "url": "java.sun.com/javase/6/docs/api/" }, { "name":
"Jon Doe", "url": "code.google.com/p/jaql/" }, { "name": "Jane Doe", "url":
"www.cnn.com" } ]
+[ { "name": "Jon Doe", "url": "java.sun.com/javase/6/docs/api/" }, { "name":
"Jon Doe", "url": "code.google.com/p/jaql/" }, { "name": "Jane Doe", "url":
"www.cnn.com" } ]
+
+# 17:33:31 >
+# 17:33:31 > "Done."
+# 17:33:31 >
+
diff --git a/monetdb5/extras/jaql/Tests/join01.mal
b/monetdb5/extras/jaql/Tests/join01.mal
--- a/monetdb5/extras/jaql/Tests/join01.mal
+++ b/monetdb5/extras/jaql/Tests/join01.mal
@@ -1,6 +1,6 @@
# from the docs
-jaql.x("users = [ {\"name\": "Jon Doe", \"password\": \"asdf1234\", \"id\":
1}, {\"name\": "Jane Doe", \"password\": \"qwertyui\", \"id\": 2}, {\"name\":
"Max Mustermann", \"password\": \"q1w2e3r4\", \"id\": 3} ];");
+jaql.x("users = [ {\"name\": \"Jon Doe\", \"password\": \"asdf1234\", \"id\":
1}, {\"name\": \"Jane Doe\", \"password\": \"qwertyui\", \"id\": 2}, {\"name\":
\"Max Mustermann\", \"password\": \"q1w2e3r4\", \"id\": 3} ];");
jaql.x("pages = [ {\"userid\": 1, \"url\":\"code.google.com/p/jaql/\"},
{\"userid\": 2, \"url\":\"www.cnn.com\"}, {\"userid\": 1,
\"url\":\"java.sun.com/javase/6/docs/api/\"} ];");
jaql.x("join users, pages where users.id == pages.userid into {users.name,
pages.*};");
diff --git a/monetdb5/extras/jaql/jaql.c b/monetdb5/extras/jaql/jaql.c
--- a/monetdb5/extras/jaql/jaql.c
+++ b/monetdb5/extras/jaql/jaql.c
@@ -326,15 +326,29 @@ _check_exp_equals_only(tree *t)
if (t == NULL)
return NULL;
- if (t->type == j_pred && t->tval2->type == j_comp && (
- t->tval2->cval != j_and &&
- t->tval2->cval != j_or &&
- t->tval2->cval != j_equals))
- {
- res = GDKzalloc(sizeof(tree));
- res->type = j_error;
- res->sval = GDKstrdup("join: only equality tests are allowed");
- return res;
+ if (t->type == j_pred && t->tval2->type == j_comp) {
+ if (t->tval2->cval != j_and && t->tval2->cval != j_equals) {
+ res = GDKzalloc(sizeof(tree));
+ res->type = j_error;
+ res->sval = GDKstrdup("join: only (conjunctions of)
equality "
+ "tests are allowed");
+ return res;
+ }
+ if (t->tval2->cval == j_equals) {
+ if (t->tval1->type != j_var || t->tval3->type != j_var)
{
+ res = GDKzalloc(sizeof(tree));
+ res->type = j_error;
+ res->sval = GDKstrdup("join: equality tests
must be between "
+ "two variables");
+ return res;
+ }
+ if (strcmp(t->tval1->sval, t->tval3->sval) == 0) {
+ res = GDKzalloc(sizeof(tree));
+ res->type = j_error;
+ res->sval = GDKstrdup("join: self-joins not
allowed");
+ return res;
+ }
+ }
}
if ((res = _check_exp_equals_only(t->tval1)) != NULL)
@@ -347,6 +361,30 @@ _check_exp_equals_only(tree *t)
return NULL;
}
+static tree *
+_extract_equals(tree *t)
+{
+ tree *p, *q;
+
+ assert(t->type == j_pred);
+ assert(t->tval2->type == j_comp);
+
+ if (t->tval2->cval == j_equals)
+ return t;
+
+ assert(t->tval2->cval == j_and);
+
+ p = q = _extract_equals(t->tval1);
+ while (p->next != NULL)
+ p = p->next;
+ p->next = _extract_equals(t->tval3);
+
+ t->tval1 = t->tval3 = t->next = NULL;
+ freetree(t);
+
+ return q;
+}
+
/* create a join operation over 2 or more inputs, applying predicates,
* producing output defined by tmpl */
tree *
@@ -383,9 +421,40 @@ make_jaql_join(tree *inputs, tree *pred,
if ((res = _check_exp_var("join", vars, tmpl)) != NULL)
return res;
+ /* JAQL defines that only conjunctions of equality expressions may
+ * be used (and + ==), where self-joins are disallowed */
if ((res = _check_exp_equals_only(pred)) != NULL)
return res;
+ /* JAQL defines that each of the inputs must be linked through a
+ * join path, collect all equality tests and put them in a simple
+ * list */
+ pred = _extract_equals(pred);
+ for (i = 0; vars[i] != NULL; i++) {
+ for (res = pred; res != NULL; res = res->next) {
+ if (strcmp(vars[i], res->tval1->sval) == 0 ||
+ strcmp(vars[i], res->tval3->sval) == 0)
+ {
+ vars[i] = "";
+ break;
+ }
+ }
+ if (vars[i][0] != '\0') {
+ char buf[128];
+ res = GDKzalloc(sizeof(tree));
+ res->type = j_error;
+ snprintf(buf, sizeof(buf), "join: input not referenced "
+ "in where: %s", vars[i]);
+ res->sval = GDKstrdup(buf);
+ freetree(inputs);
+ freetree(pred);
+ freetree(tmpl);
+ return res;
+ }
+ }
+
+ /* FIXME: we skip the graph/path check for now */
+
res = GDKzalloc(sizeof(tree));
res->type = j_join;
res->tval1 = inputs;
@@ -1035,6 +1104,9 @@ printtree(tree *t, int level, char op)
case j_divide:
printf("/ ");
break;
+ case j_cinvalid:
+ printf("<<invalid compare
node>>");
+ break;
}
break;
case j_join_input:
@@ -1064,12 +1136,16 @@ printtree(tree *t, int level, char op)
printf(", ");
printtree(t->tval3, level + step, op);
printf(") ");
+ if (t->next != NULL)
+ printf("&& ");
} else {
printf("( ");
printtree(t->tval1, level + step, op);
printtree(t->tval2, level + step, op);
printtree(t->tval3, level + step, op);
printf(") ");
+ if (t->next != NULL)
+ printf("and ");
}
break;
case j_operation:
@@ -1140,6 +1216,9 @@ printtree(tree *t, int level, char op)
printf("!%s\n", t->sval);
}
break;
+ case j_invalid:
+ printf("<<invalid tree node>>");
+ break;
}
if (t != NULL)
t = t->next;
diff --git a/monetdb5/extras/jaql/jaql.h b/monetdb5/extras/jaql/jaql.h
--- a/monetdb5/extras/jaql/jaql.h
+++ b/monetdb5/extras/jaql/jaql.h
@@ -56,6 +56,7 @@ typedef struct _jc {
} jc;
enum treetype {
+ j_invalid,
j_output_var,
j_output,
j_json,
@@ -84,6 +85,7 @@ enum treetype {
};
enum comptype {
+ j_cinvalid,
j_equals,
j_nequal,
j_greater,
@@ -142,6 +144,7 @@ tree *make_double(double d);
tree *make_string(char *s);
tree *make_bool(char b);
void printtree (tree *t, int level, char op);
+void freetree(tree *t);
jaql_export str JAQLexecute(Client cntxt, MalBlkPtr mb, MalStkPtr stk,
InstrPtr pci);
diff --git a/monetdb5/extras/jaql/jaqlgencode.c
b/monetdb5/extras/jaql/jaqlgencode.c
--- a/monetdb5/extras/jaql/jaqlgencode.c
+++ b/monetdb5/extras/jaql/jaqlgencode.c
@@ -21,7 +21,28 @@
#include "jaqlgencode.h"
#include "opt_prelude.h"
+typedef struct _json_var {
+ const char *name;
+ char preserve;
+ int j1;
+ int j2;
+ int j3;
_______________________________________________
Checkin-list mailing list
[email protected]
http://mail.monetdb.org/mailman/listinfo/checkin-list
