[Swift-commit] r3334 - provenancedb
noreply at svn.ci.uchicago.edu
noreply at svn.ci.uchicago.edu
Mon May 24 16:28:39 CDT 2010
Author: lgadelha
Date: 2010-05-24 16:28:38 -0500 (Mon, 24 May 2010)
New Revision: 3334
Modified:
provenancedb/prov-init.sql
Log:
Creation of annotation table.
Modified: provenancedb/prov-init.sql
===================================================================
--- provenancedb/prov-init.sql 2010-05-21 20:15:34 UTC (rev 3333)
+++ provenancedb/prov-init.sql 2010-05-24 21:28:38 UTC (rev 3334)
@@ -16,7 +16,8 @@
DROP TABLE extrainfo;
DROP TABLE createarray;
DROP TABLE createarray_member;
-DROP TABLE arrayrange;
+DROP TABLE array_range;
+DROP TABLE annotations;
-- associates each process with its containing workflow
@@ -24,14 +25,14 @@
-- in which case this looks very much like a compound/app
-- containment?
CREATE TABLE processes_in_workflows
- (workflow_id char(128),
- process_id char(128)
+ (workflow_id char(256),
+ process_id char(256)
);
-- processes gives information about each process (in the OPM sense)
-- it is augmented by information in other tables
CREATE TABLE processes
- (id char(128) PRIMARY KEY, -- a uri
+ (id char(256) PRIMARY KEY, -- a uri
type char(16) -- specifies the type of process. for any type, it
-- must be the case that the specific type table
-- has an entry for this process.
@@ -43,12 +44,12 @@
-- each execute is identified by a unique URI. other information from
-- swift logs is also stored here. an execute is an OPM process.
CREATE TABLE executes
- (id char(128) PRIMARY KEY, -- actually foreign key to processes
+ (id char(256) PRIMARY KEY, -- actually foreign key to processes
starttime numeric,
duration numeric,
- finalstate char(128),
- app char(128),
- scratch char(128)
+ finalstate char(256),
+ app char(256),
+ scratch char(256)
);
-- this gives information about each execute2, which is an attempt to
@@ -56,12 +57,12 @@
-- information such as wrapper logs
CREATE TABLE execute2s
- (id char(128) PRIMARY KEY,
- execute_id, -- secondary key to executes and processes tables
+ (id char(256) PRIMARY KEY,
+ execute_id char(256), -- secondary key to executes and processes tables
starttime numeric,
duration numeric,
- finalstate char(128),
- site char(128)
+ finalstate char(256),
+ site char(256)
);
-- dataset_usage records usage relationships between processes and datasets;
@@ -73,11 +74,11 @@
-- dataset_id for common queries? maybe add arbitrary ID for sake of it?
CREATE TABLE dataset_usage
- (process_id char(128), -- foreign key but not enforced because maybe process
+ (process_id char(256), -- foreign key but not enforced because maybe process
-- doesn't exist at time. same type as processes.id
direction char(1), -- I or O for input or output
- dataset_id char(128), -- this will perhaps key against dataset table
- param_name char(128) -- the name of the parameter in this execute that
+ dataset_id char(256), -- this will perhaps key against dataset table
+ param_name char(256) -- the name of the parameter in this execute that
-- this dataset was bound to. sometimes this must
-- be contrived (for example, in positional varargs)
);
@@ -91,8 +92,8 @@
-- TODO primary key should be execute_id
CREATE TABLE invocation_procedure_names
- (execute_id char(128),
- procedure_name char(128)
+ (execute_id char(256),
+ procedure_name char(256)
);
@@ -106,8 +107,8 @@
-- a containment hierarchy. The relationship (such as array index or
-- structure member name) should also be stored in this table.
CREATE TABLE dataset_containment
- ( outer_dataset_id char(128),
- inner_dataset_id char(128)
+ ( outer_dataset_id char(256),
+ inner_dataset_id char(256)
);
@@ -117,8 +118,8 @@
-- TODO dataset_id should be primary key
CREATE TABLE dataset_filenames
- ( dataset_id char(128),
- filename char(128)
+ ( dataset_id char(256),
+ filename char(256)
);
-- dataset_values stores the value for each dataset which is known to have
@@ -127,8 +128,8 @@
-- example) SQL numerical operations should not be expected to work, even
-- though the user knows that a particular dataset stores a numeric value.
CREATE TABLE dataset_values
- ( dataset_id char(128), -- should be primary key
- value char(128)
+ ( dataset_id char(256), -- should be primary key
+ value char(256)
);
-- The above dataset_* tables are the original containment representation
@@ -138,21 +139,21 @@
-- It is unclear which is the better representation.
CREATE TABLE createarray
- ( array_id char(128)
+ ( array_id char(256)
);
CREATE TABLE createarray_member
- ( array_id char(128),
- ix char(128),
- member_id char(128)
+ ( array_id char(256),
+ ix char(256),
+ member_id char(256)
);
-- TODO step
CREATE TABLE array_range
- ( array_id char(128),
- from_id char(128),
- to_id char(128),
- step_id char(128) -- nullable, if step is unspecified
+ ( array_id char(256),
+ from_id char(256),
+ to_id char(256),
+ step_id char(256) -- nullable, if step is unspecified
);
-- known_workflows stores some information about each workflow log that has
@@ -160,27 +161,33 @@
-- status.
CREATE TABLE known_workflows
(
- workflow_id char(128),
- workflow_log_filename char(128),
- version char(128),
- importstatus char(128)
+ workflow_id char(256),
+ workflow_log_filename char(256),
+ version char(256),
+ importstatus char(256)
);
-- workflow_events stores the start time and duration for each workflow
-- that has been successfully imported.
CREATE TABLE workflow_events
- ( workflow_id char(128),
+ ( workflow_id char(256),
starttime numeric,
duration numeric
);
-- extrainfo stores lines generated by the SWIFT_EXTRA_INFO feature
CREATE TABLE extrainfo
- ( execute2id char(128),
+ ( execute2id char(256),
extrainfo char(1024)
);
+-- annotations
+CREATE TABLE annotations
+ ( id char(256), -- either dataset_id, process_id, or workflow_id
+ name char(256),
+ value char(256)
+ );
-- this GRANT does not work for sqlite; you'll get a syntax error but
-- ignore it, as it is not needed in sqlite
More information about the Swift-commit
mailing list