Release 0.24.0

Improvements in `CDB_JenksBins`

.travis.yml

@@ -1,43 +1,23 @@
 language: c
-dist: precise
+dist: trusty
 sudo: required
 
-addons:
-  postgresql: 9.5
+env:
+  - PGUSER=postgres PGDATABASE=postgres
+
+env:
+    - POSTGRESQL_VERSION="9.5", POSTGIS_VERSION="2.4"
+    - POSTGRESQL_VERSION="10", POSTGIS_VERSION="2.4"
 
 before_install:
-  # Add custom PPAs from cartodb
-  - sudo add-apt-repository -y ppa:cartodb/postgresql-9.5
-  - sudo add-apt-repository -y ppa:cartodb/gis
-  - sudo add-apt-repository -y ppa:cartodb/gis-testing
-  - sudo apt-get update
-
-  # Force instalation of libgeos-3.5.0 (presumably needed because of existing version of postgis)
-  - sudo apt-get -y install libgeos-3.5.0=3.5.0-1cdb2
-
-  # Install postgres db and build deps
-  - sudo /etc/init.d/postgresql stop # stop travis default instance
-  - sudo apt-get -y remove --purge postgresql-9.1
-  - sudo apt-get -y remove --purge postgresql-9.2
-  - sudo apt-get -y remove --purge postgresql-9.3
-  - sudo apt-get -y remove --purge postgresql-9.4
-  - sudo apt-get -y remove --purge postgresql-9.5
-  - sudo rm -rf /var/lib/postgresql/
-  - sudo rm -rf /var/log/postgresql/
-  - sudo rm -rf /etc/postgresql/
-  - sudo apt-get -y remove --purge postgis-2.2
-  - sudo apt-get -y autoremove
-
-  - sudo apt-get -y install postgresql-9.5=9.5.2-3cdb3
-  - sudo apt-get -y install postgresql-server-dev-9.5=9.5.2-3cdb3
-  - sudo apt-get -y install postgresql-plpython-9.5=9.5.2-3cdb3
-  - sudo apt-get -y install postgresql-9.5-postgis-scripts=2.2.2.0-cdb2
-  - sudo apt-get -y install postgresql-9.5-postgis-2.2=2.2.2.0-cdb2
-
-  # configure it to accept local connections from postgres
-  - echo -e "# TYPE  DATABASE        USER            ADDRESS                 METHOD \nlocal   all             postgres                                trust\nlocal   all             all                                     trust\nhost    all             all             127.0.0.1/32            trust" \
-    | sudo tee /etc/postgresql/9.5/main/pg_hba.conf
-  - sudo /etc/init.d/postgresql restart 9.5
+  - export POSTGIS_PACKAGE_VERSION=$(apt list postgis -a 2>/dev/null | grep $POSTGIS_VERSION | awk '{ print $2 }')
+  - sudo service postgresql stop;
+  - sudo apt-get install -y --allow-unauthenticated --no-install-recommends --no-install-suggests postgresql-$POSTGRESQL_VERSION postgresql-client-$POSTGRESQL_VERSION postgresql-server-dev-$POSTGRESQL_VERSION postgis=$POSTGIS_PACKAGE_VERSION postgresql-$POSTGRESQL_VERSION-postgis-scripts postgresql-$POSTGRESQL_VERSION-postgis-$POSTGIS_VERSION postgresql-plpython-$POSTGRESQL_VERSION
+  - echo -e "# TYPE  DATABASE        USER            ADDRESS                 METHOD \nlocal   all             postgres                                trust\nlocal   all             all                                     trust\nhost    all             all             127.0.0.1/32            trust"  | sudo tee /etc/postgresql/$POSTGRESQL_VERSION/main/pg_hba.conf
+  - export PGPORT=`grep ^port /etc/postgresql/$POSTGRESQL_VERSION/main/postgresql.conf | awk '{print $3}'`
+  - export PGUSER=postgres
+  - export PGDATABASE=postgres
+  - sudo service postgresql restart $POSTGRESQL_VERSION;
 
 script:
   - make

Makefile

@@ -1,7 +1,7 @@
 # cartodb/Makefile
 
 EXTENSION = cartodb
-EXTVERSION = 0.23.2
+EXTVERSION = 0.24.0
 
 SED = sed
 AWK = awk
@@ -91,6 +91,7 @@ UPGRADABLE = \
   0.23.0 \
   0.23.1 \
   0.23.2 \
+  0.24.0 \
   $(EXTVERSION)dev \
   $(EXTVERSION)next \
   $(END)

NEWS.md

@@ -1,3 +1,14 @@
+0.24.0 (2018-09-13)
+* Travis: Test with PostgreSQL 9.5 and 10.
+* _cdb_estimated_extent: Fix bug with ST_EstimatedExtent interaction.
+* Improvements in `CDB_JenksBins`.
+  * Now it ignores NULLs.
+  * No longer puts the same value in multiple categories.
+  * Removes all limits related to size.
+  * If not set, the number of iterations done is based now on the size of the array.
+  * Fixed multiple bugs.
+  * The internal function `CDB_JenksBinsIteration` has changed its signature.
+
 0.23.2 (2018-07-19)
 * Fix `CDB_QueryTablesText` with parenthesized queries (#335)
 

carto-package.json

@@ -0,0 +1,11 @@
+{
+    "name": "carto_postgresql_ext",
+    "current_version": {
+        "requires": {
+            "postgresql": ">=9.5.0",
+            "postgis": ">=2.2.0.0"
+        },
+        "works_with": {
+        }
+    }
+}

scripts-available/CDB_JenksBins.sql

@@ -10,212 +10,337 @@
 --
 -- @param invert Optional wheter to return the top of each bin (default)
 --               or the bottom. BOOLEAN, default=FALSE.
---  
+--
 --
 
+CREATE OR REPLACE FUNCTION CDB_JenksBins(in_array NUMERIC[], breaks INT, iterations INT DEFAULT 0, invert BOOLEAN DEFAULT FALSE)
+RETURNS NUMERIC[] as
+$$
+DECLARE
+    in_matrix NUMERIC[][];
+    in_unique_count BIGINT;
 
-CREATE OR REPLACE FUNCTION CDB_JenksBins ( in_array NUMERIC[], breaks INT, iterations INT DEFAULT 5, invert BOOLEAN DEFAULT FALSE) RETURNS NUMERIC[] as $$ 
-DECLARE 
-    element_count INT4; 
+    shuffles INT;
     arr_mean NUMERIC;
+    sdam NUMERIC;
+
+    i INT;
     bot INT;
     top INT;
+
     tops INT[];
     classes INT[][];
-    i INT := 1; j INT := 1; 
+    j INT := 1;
     curr_result NUMERIC[];
     best_result NUMERIC[];
     seedtarget TEXT;
-    quant NUMERIC[];
-    shuffles INT;
+
 BEGIN
-    -- get the total size of our row
-    element_count := array_length(in_array, 1); --array_upper(in_array, 1) - array_lower(in_array, 1); 
-    -- ensure the ordering of in_array
-    SELECT array_agg(e) INTO in_array FROM (SELECT unnest(in_array) e ORDER BY e) x;
-    -- stop if no rows 
-    IF element_count IS NULL THEN  
-        RETURN NULL; 
-    END IF; 
-    -- stop if our breaks are more than our input array size
-    IF element_count < breaks THEN  
-        RETURN in_array; 
-    END IF; 
+    -- We clean the input array (remove NULLs) and create 2 arrays
+    -- [1] contains the unique values in in_array
+    -- [2] contains the number of appearances of those unique values
+    SELECT ARRAY[array_agg(value), array_agg(count)] FROM
+    (
+        SELECT value, count(1)::numeric as count
+        FROM  unnest(in_array) AS value
+        WHERE value is NOT NULL
+        GROUP BY value
+        ORDER BY value
+    ) __clean_array_q INTO in_matrix;
 
-    shuffles := LEAST(GREATEST(floor(2500000.0/(element_count::float*iterations::float)), 1), 750)::int;
-    -- get our mean value
-    SELECT avg(v) INTO arr_mean FROM (  SELECT unnest(in_array) as v ) x; 
+    -- Get the number of unique values
+    in_unique_count := array_length(in_matrix[1:1], 2);
 
-    -- assume best is actually Quantile
-    SELECT CDB_QuantileBins(in_array, breaks) INTO quant;
-
-    -- if data is very very large, just return quant and be done
-    IF element_count > 5000000 THEN
-        RETURN quant;
+    IF in_unique_count IS NULL THEN
+        RETURN NULL;
     END IF;
 
-    -- change quant into bottom, top markers
-    LOOP 
-        IF i = 1 THEN 
-            bot = 1;
-        ELSE 
-            -- use last top to find this bot
-            bot = top+1;
-        END IF;
-        IF i = breaks THEN
-            top = element_count;
+    IF in_unique_count <= breaks THEN
+        -- There isn't enough distinct values for the requested breaks
+        RETURN ARRAY(Select unnest(in_matrix[1:1])) _a;
+    END IF;
+
+    -- If not declated explicitly we iterate based on the length of the array
+    IF iterations < 1 THEN
+        -- This is based on a 'looks fine' heuristic
+        iterations := log(in_unique_count)::integer + 1;
+    END IF;
+
+    -- We set the number of shuffles per iteration as the number of unique values but
+    -- this is just another 'looks fine' heuristic
+    shuffles := in_unique_count;
+
+    -- Get the mean value of the whole vector (already ignores NULLs)
+    SELECT avg(v) INTO arr_mean FROM ( SELECT unnest(in_array) as v ) x;
+
+    -- Calculate the sum of squared deviations from the array mean (SDAM).
+    SELECT sum(((arr_mean - v)^2) * w) INTO sdam FROM (
+        SELECT unnest(in_matrix[1:1]) as v, unnest(in_matrix[2:2]) as w
+        ) x;
+
+    -- To start, we create ranges with approximately the same amount of different values
+    top := 0;
+    i := 1;
+    LOOP
+        bot := top + 1;
+        top := ROUND(i * in_unique_count::numeric / breaks::NUMERIC);
+
+        IF i = 1 THEN
+            classes = ARRAY[ARRAY[bot,top]];
         ELSE
-            SELECT count(*) INTO top FROM ( SELECT unnest(in_array) as v) x WHERE v <= quant[i];
-        END IF;
-        IF i = 1 THEN 
-            classes = ARRAY[ARRAY[bot,top]]; 
-        ELSE 
-            classes = ARRAY_CAT(classes,ARRAY[bot,top]); 
+            classes = ARRAY_CAT(classes, ARRAY[bot,top]);
         END IF;
+
+        i := i + 1;
         IF i > breaks THEN EXIT; END IF;
-        i = i+1;
     END LOOP;
 
-    best_result = CDB_JenksBinsIteration( in_array, breaks, classes, invert, element_count, arr_mean, shuffles);
+    best_result = CDB_JenksBinsIteration(in_matrix, breaks, classes, invert, sdam, shuffles);
 
     --set the seed so we can ensure the same results
     SELECT setseed(0.4567) INTO seedtarget;
     --loop through random starting positions
     LOOP
-        IF j > iterations-1 THEN  EXIT;  END IF;  
+        IF j > iterations-1 THEN  EXIT;  END IF;
         i = 1;
-        tops = ARRAY[element_count];
+        tops = ARRAY[in_unique_count];
         LOOP
-            IF i = breaks THEN  EXIT;  END IF;  
-            SELECT array_agg(distinct e) INTO tops FROM (SELECT unnest(array_cat(tops, ARRAY[floor(random()*element_count::float)::int])) as e ORDER BY e) x WHERE e != 1;
-            i = array_length(tops, 1); 
-        END LOOP; 
+            IF i = breaks THEN  EXIT;  END IF;
+            SELECT array_agg(distinct e) INTO tops FROM (
+                SELECT unnest(array_cat(tops, ARRAY[trunc(random() * in_unique_count::float8)::int + 1])) as e ORDER BY e
+                ) x;
+            i = array_length(tops, 1);
+        END LOOP;
+        top := 0;
         i = 1;
-        LOOP  
-            IF i > breaks THEN  EXIT;  END IF;  
-            IF i = 1 THEN
-                bot = 1;
-            ELSE
-                bot = top+1;
-            END IF;
+        LOOP
+            bot := top + 1;
             top = tops[i];
-            IF i = 1 THEN 
-                classes = ARRAY[ARRAY[bot,top]]; 
-            ELSE 
-                classes = ARRAY_CAT(classes,ARRAY[bot,top]); 
+            IF i = 1 THEN
+                classes = ARRAY[ARRAY[bot,top]];
+            ELSE
+                classes = ARRAY_CAT(classes, ARRAY[bot,top]);
             END IF;
-            i := i+1; 
-        END LOOP; 
-        curr_result = CDB_JenksBinsIteration( in_array, breaks, classes, invert, element_count, arr_mean, shuffles);
+
+            i := i+1;
+            IF i > breaks THEN EXIT; END IF;
+        END LOOP;
+
+        curr_result = CDB_JenksBinsIteration(in_matrix, breaks, classes, invert, sdam, shuffles);
 
         IF curr_result[1] > best_result[1] THEN
             best_result = curr_result;
-            j = j-1; -- if we found a better result, add one more search
         END IF;
+
         j = j+1;
     END LOOP;
 
     RETURN (best_result)[2:array_upper(best_result, 1)];
 END;
-$$ language plpgsql VOLATILE PARALLEL RESTRICTED;
-
+$$ LANGUAGE PLPGSQL IMMUTABLE PARALLEL RESTRICTED;
 
 
 --
 -- Perform a single iteration of the Jenks classification
 --
+-- Returns an array with:
+-- - First element: gvf
+-- - Second to 2+n: Category limits
+DROP FUNCTION IF EXISTS CDB_JenksBinsIteration ( in_matrix NUMERIC[], breaks INT, classes INT[], invert BOOLEAN, element_count INT4, arr_mean NUMERIC, max_search INT); -- Old signature
+
+CREATE OR REPLACE FUNCTION CDB_JenksBinsIteration ( in_matrix NUMERIC[], breaks INT, classes INT[], invert BOOLEAN, sdam NUMERIC, max_search INT DEFAULT 50) RETURNS NUMERIC[] as $$
+DECLARE
+    i INT;
+    iterations INT = 0;
 
-CREATE OR REPLACE FUNCTION CDB_JenksBinsIteration ( in_array NUMERIC[], breaks INT, classes INT[][], invert BOOLEAN, element_count INT4, arr_mean NUMERIC, max_search INT DEFAULT 50) RETURNS NUMERIC[] as $$ 
-DECLARE 
-    tmp_val numeric; 
-    new_classes int[][];
-    tmp_class int[];
-    i INT := 1; 
-    j INT := 1; 
     side INT := 2;
-    sdam numeric; 
-    gvf numeric := 0.0; 
-    new_gvf numeric; 
-    arr_gvf numeric[]; 
-    class_avg numeric; 
-    class_max_i INT; 
-    class_min_i INT; 
-    class_max numeric; 
-    class_min numeric; 
-    reply numeric[]; 
-BEGIN 
 
-    -- Calculate the sum of squared deviations from the array mean (SDAM).
-    SELECT sum((arr_mean - e)^2) INTO sdam FROM (  SELECT unnest(in_array) as e ) x; 
-    --Identify the breaks for the lowest GVF
-    LOOP   
-        i = 1; 
-        LOOP 
-            -- get our mean
-            SELECT avg(e) INTO class_avg FROM ( SELECT unnest(in_array[classes[i][1]:classes[i][2]]) as e) x;  
-            -- find the deviation
-            SELECT sum((class_avg-e)^2) INTO tmp_val FROM (   SELECT unnest(in_array[classes[i][1]:classes[i][2]]) as e  ) x;  
-            IF i = 1 THEN 
-                arr_gvf = ARRAY[tmp_val]; 
-                -- init our min/max map for later
-                class_max = arr_gvf[i];  
-                class_min = arr_gvf[i];  
-                class_min_i = 1;   
-                class_max_i = 1;  
-            ELSE 
-                arr_gvf = array_append(arr_gvf, tmp_val); 
-            END IF;
-            i := i+1;  
-            IF i > breaks THEN EXIT; END IF;  
-        END LOOP;  
-        -- calculate our new GVF
-        SELECT sdam-sum(e) INTO new_gvf FROM (  SELECT unnest(arr_gvf) as e  ) x;  
-        -- if no improvement was made, exit
-        IF new_gvf < gvf THEN EXIT; END IF; 
-        gvf = new_gvf;  
-        IF j > max_search THEN EXIT; END IF; 
-        j = j+1;
-        i = 1;  
-        LOOP  
-            --establish directionality (uppward through classes or downward)
-            IF arr_gvf[i] < class_min THEN   
-                class_min = arr_gvf[i];   
-                class_min_i = i;  
-            END IF;  
-            IF arr_gvf[i] > class_max THEN   
-                class_max = arr_gvf[i];   
-                class_max_i = i;  
-            END IF;  
-            i := i+1;  
-            IF i > breaks THEN EXIT; END IF;  
-        END LOOP;  
-        IF class_max_i > class_min_i THEN
-            class_min_i = class_max_i - 1;
+    gvf numeric := 0.0;
+    new_gvf numeric;
+    arr_gvf numeric[];
+    arr_avg numeric[];
+    class_avg numeric;
+    class_dev numeric;
+
+    class_max_i INT = 0;
+    class_min_i INT = 0;
+    dev_max numeric;
+    dev_min numeric;
+
+    best_classes INT[] = classes;
+    best_gvf numeric[];
+    best_avg numeric[];
+    move_elements INT = 1;
+
+    reply numeric[];
+
+BEGIN
+
+    -- We fill the arrays with the initial values
+    i = 0;
+    LOOP
+        IF i = breaks THEN EXIT; END IF;
+        i = i + 1;
+
+        -- Get class mean
+        SELECT (sum(v * w) / sum(w)) INTO class_avg FROM (
+            SELECT unnest(in_matrix[1:1][classes[i][1]:classes[i][2]]) as v,
+                    unnest(in_matrix[2:2][classes[i][1]:classes[i][2]]) as w
+            ) x;
+
+        -- Get class deviation
+        SELECT sum((class_avg - v)^2 * w) INTO class_dev FROM (
+            SELECT unnest(in_matrix[1:1][classes[i][1]:classes[i][2]]) as v,
+                    unnest(in_matrix[2:2][classes[i][1]:classes[i][2]]) as w
+            ) x;
+
+
+        IF i = 1 THEN
+            arr_avg = ARRAY[class_avg];
+            arr_gvf = ARRAY[class_dev];
         ELSE
-            class_min_i = class_max_i + 1;
+            arr_avg = array_append(arr_avg, class_avg);
+            arr_gvf = array_append(arr_gvf, class_dev);
         END IF;
-            --Move from higher class to a lower gid order
-            IF class_max_i > class_min_i THEN
-                classes[class_max_i][1] = classes[class_max_i][1] + 1;
-                classes[class_min_i][2] = classes[class_min_i][2] + 1;
-            ELSE -- Move from lower class UP into a higher class by gid
-                classes[class_max_i][2] = classes[class_max_i][2] - 1;
-                classes[class_min_i][1] = classes[class_min_i][1] - 1;
+    END LOOP;
+
+    -- We copy the values to avoid recalculation when a failure happens
+    best_avg = arr_avg;
+    best_gvf = arr_gvf;
+
+    iterations = 0;
+    LOOP
+        IF iterations = max_search THEN EXIT; END IF;
+        iterations = iterations + 1;
+
+        -- calculate our new GVF
+        SELECT sdam - sum(e) INTO new_gvf FROM ( SELECT unnest(arr_gvf) as e ) x;
+
+        -- Check if any improvement was made
+        IF new_gvf <= gvf THEN
+            -- If we were moving too many elements, go back and move less
+            IF move_elements <= 2 OR class_max_i = class_min_i THEN
+                EXIT;
             END IF;
-    END LOOP; 
+
+            move_elements = GREATEST(move_elements / 8, 1);
+
+            -- Rollback from saved statuses
+            classes = best_classes;
+            new_gvf = gvf;
+
+            i = class_min_i;
+            LOOP
+                arr_avg[i] = best_avg[i];
+                arr_gvf[i] = best_gvf[i];
+
+                IF i = class_max_i THEN EXIT; END IF;
+                i = i + 1;
+            END LOOP;
+        END IF;
+
+        -- We search for the classes with the min and max deviation
+        i = 1;
+        class_min_i = 1;
+        class_max_i = 1;
+        dev_max = arr_gvf[1];
+        dev_min = arr_gvf[1];
+        LOOP
+            IF i = breaks THEN EXIT; END IF;
+            i = i + 1;
+
+            IF arr_gvf[i] < dev_min THEN
+                dev_min = arr_gvf[i];
+                class_min_i = i;
+            ELSE
+                IF arr_gvf[i] > dev_max THEN
+                    dev_max = arr_gvf[i];
+                    class_max_i = i;
+                END IF;
+            END IF;
+        END LOOP;
+
+
+        -- Save best values for comparison and output
+        gvf = new_gvf;
+        best_classes = classes;
+
+        -- Limit the moved elements as to not remove everything from class_max_i
+        move_elements = LEAST(move_elements, classes[class_max_i][2] - classes[class_max_i][1]);
+
+        -- Move `move_elements` from class_max_i to class_min_i
+        IF class_min_i < class_max_i THEN
+            i := class_min_i;
+            LOOP
+                IF i = class_max_i THEN EXIT; END IF;
+                classes[i][2] = classes[i][2] + move_elements;
+                i := i + 1;
+            END LOOP;
+
+            i := class_max_i;
+            LOOP
+                IF i = class_min_i THEN EXIT; END IF;
+                classes[i][1] = classes[i][1] + move_elements;
+                i := i - 1;
+            END LOOP;
+        ELSE
+            i := class_min_i;
+            LOOP
+                IF i = class_max_i THEN EXIT; END IF;
+                classes[i][1] = classes[i][1] - move_elements;
+                i := i - 1;
+            END LOOP;
+
+            i := class_max_i;
+            LOOP
+                IF i = class_min_i THEN EXIT; END IF;
+                classes[i][2] = classes[i][2] - move_elements;
+                i := i + 1;
+            END LOOP;
+        END IF;
+
+        -- Recalculate avg and deviation ONLY for the affected classes
+        i = LEAST(class_min_i, class_max_i);
+        class_max_i = GREATEST(class_min_i, class_max_i);
+        class_min_i = i;
+        LOOP
+            SELECT (sum(v * w) / sum(w)) INTO class_avg FROM (
+                SELECT unnest(in_matrix[1:1][classes[i][1]:classes[i][2]]) as v,
+                        unnest(in_matrix[2:2][classes[i][1]:classes[i][2]]) as w
+                ) x;
+
+            SELECT sum((class_avg - v)^2 * w) INTO class_dev FROM (
+                SELECT unnest(in_matrix[1:1][classes[i][1]:classes[i][2]]) as v,
+                        unnest(in_matrix[2:2][classes[i][1]:classes[i][2]]) as w
+                ) x;
+
+            -- Save status (in case it's needed for rollback) and store the new one
+            best_avg[i] = arr_avg[i];
+            arr_avg[i] = class_avg;
+
+            best_gvf[i] = arr_gvf[i];
+            arr_gvf[i] = class_dev;
+
+            IF i = class_max_i THEN EXIT; END IF;
+            i = i + 1;
+        END LOOP;
+
+        move_elements = move_elements * 2;
+
+    END LOOP;
 
     i = 1;
-    LOOP  
+    LOOP
         IF invert = TRUE THEN
             side = 1; --default returns bottom side of breaks, invert returns top side
         END IF;
-        reply = array_append(reply, in_array[classes[i][side]]);  
-        i = i+1;  
-        IF i > breaks THEN  EXIT; END IF; 
-    END LOOP; 
-    
-    RETURN array_prepend(gvf, reply); 
+        reply = array_append(reply, unnest(in_matrix[1:1][best_classes[i][side]:best_classes[i][side]]));
+        i = i+1;
+        IF i > breaks THEN  EXIT; END IF;
+    END LOOP;
 
-END; 
-$$ language plpgsql IMMUTABLE PARALLEL SAFE;
+    reply = array_prepend(gvf, reply);
+    RETURN reply;
 
+END;
+$$ LANGUAGE PLPGSQL IMMUTABLE PARALLEL SAFE;

scripts-available/CDB_Overviews.sql

@@ -93,11 +93,8 @@ AS $$
       table_id.schema_name, table_id.table_name, 'the_geom_webmercator'
     );
 
-    BEGIN
-      EXECUTE ext_query INTO ext;
-    EXCEPTION
-        -- This is the typical ERROR: stats for "mytable" do not exist
-        WHEN internal_error THEN
+    EXECUTE ext_query INTO ext;
+    IF ext IS NULL THEN
           -- Get stats and execute again
           EXECUTE format('ANALYZE %1$s', reloid);
 
@@ -107,7 +104,7 @@ AS $$
           END IF;
 
           EXECUTE ext_query INTO ext;
-    END;
+    END IF;
 
     RETURN ext;
   END;

scripts-available/CDB_TableMetadata.sql

@@ -7,7 +7,7 @@ CREATE TABLE IF NOT EXISTS
 
 CREATE OR REPLACE VIEW public.CDB_TableMetadata_Text AS
        SELECT FORMAT('%I.%I', n.nspname::text, c.relname::text) tabname, updated_at
-       FROM public.CDB_TableMetadata, pg_catalog.pg_class c
+       FROM public.CDB_TableMetadata m JOIN pg_catalog.pg_class c ON m.tabname::oid = c.oid
        LEFT JOIN pg_catalog.pg_namespace n ON c.relnamespace = n.oid;
 
 -- No one can see this

test/CDB_GreatCircle_expect

@@ -1,2 +1,2 @@
-LINESTRING(4.259 55.858,5.6692453115051 56.0150275120673,7.10720375678704 56.157400475677,8.5718366560563 56.2842986378254,10.0619272412891 56.3949153508462,11.5760785994189 56.4884642014437,13.1127142001617 56.564185934303,14.6700812655504 56.6213555706215,16.2462571744128 56.6592896061102,17.8391590143095 56.6773531596105,19.4465562981665 56.6749669334121,21.0660867567155 56.6516138405427,22.6952750058883 56.6068451534252,24.3315537765309 56.540286032869,25.9722872888145 56.4516403065472,27.6147962622065 56.3406943817481,29.2563839799455 56.207320197769,30.8943627796619 56.0514771479657,32.5260803224591 55.8732129290618,34.1489450028345 55.6726633044968,35.7604499005266 55.4500507979281,37.3581947399686 55.2056823610616,38.9399054089486 54.9399460854786,40.5034506895044 54.6533070499613,42.0468559644411 54.3463024122038,43.5683137754523 54.0195358662507,45.066191217402 53.673671594382,46.5390342525062 53.3094278446298,47.9855691138079 52.9275702630659,49.4047010366934 52.5289051040742,50.7955106088955 52.1142724327331,52.1572480633875 51.6845394219955,53.4893258557794 51.2405938343407,54.7913098701049 50.7833377637432,56.0629095865715 50.3136816997865,57.3039675245588 49.8325389621027,58.5144482465496 49.3408205404538,59.6944271762829 48.8394303639846,60.8440794494795 48.329261012675,61.9636689799149 47.811189874886,63.0535378889196 47.2860757471582,64.1140964137264 46.7547558660385,65.1458133802427 46.2180433565989,66.1492072992903 45.676725078361,67.1248381223566 45.131559846414,68.0732996734468 44.583277003498,68.9952127576034 44.0325753175597,69.8912189338183 43.4801221786776,70.7619749300985 42.9265530691612,71.6081476710291 42.3724712809595,72.4304098829428 41.8184478551838,73.2294362384225 41.2650217194684,74.0059 40.7127)
-LINESTRING(4.259 55.858,4.96060044865294 55.9382939511593,5.6692453115051 56.0150275120673,6.38482117645567 56.0880973218335,7.10720375678705 56.157400475677,7.83625773770865 56.2228347173136,8.5718366560563 56.2842986378254,9.31378281572326 56.3416918804739,10.0619272412891 56.3949153508462,10.8160896721679 56.4438714316548,11.5760785994189 56.4884642014437,12.3416913471456 56.528599656387,13.1127142001617 56.5641859343031,13.8889225793161 56.5951335399513,14.6700812655504 56.6213555706215,15.4559446734179 56.6427679409819,16.2462571744128 56.6592896061102,17.0407534700619 56.6708427815999,17.8391590143095 56.6773531596105,18.6411904842936 56.6787501197174,19.4465562981665 56.6749669334121,20.2549571781681 56.6659409611101,21.0660867567155 56.6516138405428,21.8796322228404 56.6319316654367,22.6952750058883 56.6068451534252,23.5126914929996 56.5763098021872,24.3315537765309 56.5402860328691,25.1515304272452 56.4987393199198,25.9722872888145 56.4516403065472,26.7934882889404 56.3989649050969,27.6147962622065 56.3406943817482,28.4358737796488 56.2768154250305,29.2563839799456 56.207320197769,30.0759913971174 56.1322063721813,30.8943627796619 56.0514771479657,31.7111678961496 55.9651412533344,32.5260803224592 55.8732129290618,33.3387782060384 55.7757118957345,34.1489450028345 55.6726633044969,34.9562701828379 55.5640976716962,35.7604499005266 55.4500507979282,36.5611876268714 55.3305636720814,37.3581947399687 55.2056823610617,38.1511910717861 55.0754578859583,38.9399054089486 54.9399460854787,39.7240759459355 54.7992074675415,40.5034506895044 54.6533070499613,41.277787813601 54.5023141912026,42.0468559644411 54.3463024122039,42.8104345158644 54.1853492102971,43.5683137754524 54.0195358662508,44.3202951422663 53.8489472454711,45.066191217402 53.6736715943821,45.8058258688602 53.4938003329924,46.5390342525062 53.3094278446299,47.2656627911282 53.1206512637978,47.985569113808 52.9275702630661,48.6986219579803 52.7302868398744,49.4047010366934 52.5289051040743,50.1036968736777 52.3235310669909,50.7955106088955 52.1142724327332,51.4800537772815 51.9012383924278,52.1572480633875 51.6845394219956,52.8270250346284 51.4642870840378,53.4893258557795 51.2405938343408,54.1441009873167 51.0135728334539,54.791309870105 50.7833377637434,55.4309205988438 50.5500026522693,56.0629095865715 50.3136816997866,56.6872612224038 50.0744891161178,57.3039675245588 49.8325389621028,57.9130277905821 49.5879449982851,58.5144482465496 49.3408205404539,59.1082416968843 49.091278322122,59.6944271762829 48.8394303639847,60.2730296051101 48.5853878503721,60.8440794494795 48.3292610126751,61.40761238711 48.0711590197009,61.9636689799149 47.8111898748862,62.5122943541616 47.5494603202768,63.0535378889195 47.2860757471584,63.5874529134047 47.0211401132109,64.1140964137264 46.7547558660387,64.6335287494427 46.4870238729191,65.1458133802426 46.2180433565991,65.6510166029904 45.9479118369597,66.1492072992903 45.6767250783612,66.6404566936622 45.4045770424768,67.1248381223566 45.1315598464143,67.6024268127789 44.8577637259253,68.0732996734467 44.5832770034983,68.5375350943572 44.3081860611283,68.9952127576034 44.0325753175599,69.4464134580461 43.7565272097974,69.8912189338183 43.4801221786779,70.3297117064144 43.2034386583077,70.7619749300985 42.9265530691615,71.1880922503468 42.6495398146493,71.6081476710291 42.3724712809597,72.0222254300221 42.0954178399905,72.4304098829428 41.8184478551841,72.8327853946822 41.5416276900883,73.2294362384225 41.2650217194687,73.6204465018146 40.9886923428039,74.0059 40.7127)
+LINESTRING(4.259 55.858,5.53349240387128 56.0006659105918,6.81698919498694 56.130094578525,8.10870381314147 56.2461317260662,9.40781156033233 56.3486370295466,10.7134527044527 56.4374849223869,12.0247359780093 56.5125653297878,13.3407424424749 56.573784325367,14.660529681225 56.6210647008095,15.9831362768927 56.654346440595,17.307586522649 56.67358709506,18.6328953115992 56.6787620464102,19.9580731443722 56.6698646638042,21.282131192215 56.6469063452276,22.6040863516019 56.6099164455407,23.922966226566 56.5589420917603,25.2378139766594 56.4940478882858,26.5476929715805 56.4153155163602,27.8516911979552 56.3228432335229,29.1489253693643 56.2167452801302,30.4385446972665 56.0971512011646,31.7197342877491 55.9642050924945,32.9917181368037 55.8180647814723,34.2537617048216 55.6589009522625,35.5051740589896 55.4868962265697,36.7453095800251 55.3022442104976,37.9735692370026 55.1051485181267,39.1894014407465 54.8958217820713,40.3923024922398 54.6744846607816,41.5818166476476 54.4413648517294,42.757535825811 54.1966961188706,43.919098987406 53.9407173419567,45.0661912174019 53.673671594382,46.198542544017 53.3958052553427,47.3159265281308 53.1073671611612,48.4181586571351 52.8086077997244,49.5050945765883 52.499778551104,50.5766281918714 52.1811309766006,51.6326896704254 51.8529161576737,52.673243373185 51.5153840855177,53.6982857415906 51.1687831014009,54.7078431641625 50.8133593873253,55.7019698441171 50.4493565060761,56.6807456869812 50.0770149893128,57.6442742246566 49.6965719720156,58.5926805899637 49.3082608713202,59.5261095533829 48.9123111075629,60.4447236315382 48.5089478652008,61.3487012749643 48.0983918911668,62.2382351408597 47.6808593281578,63.1135304548766 47.2565615803358,63.9748034645285 46.8257052089336,64.822279985501 46.3884918552974,65.6561940410346 45.9451181889661,66.476786593589 45.4957758784676,67.284304367196 45.0406515826125,68.0789987582454 44.5799269601738,68.8611248319107 44.1137786959568,69.6309404010034 43.6423785413868,70.388705183725 43.1658933678633,71.1346800365587 42.6844852312539,71.8691262583921 42.1983114460249,72.5923049618788 41.7075246676227,73.3044765080245 41.2122729818388,74.0059 40.7127)
+LINESTRING(4.259 55.858,4.89507305967085 55.930977446384,5.53349240387128 56.0006659105918,6.17416348361598 56.0670448594645,6.81698919498694 56.130094578525,7.46186995983655 56.1897961993418,8.10870381314147 56.2461317260662,8.75738649688733 56.2990840610623,9.40781156033233 56.3486370295466,10.0598704664666 56.3947754031591,10.7134527044527 56.4374849223869,11.3684459078018 56.4767523177655,12.0247359780093 56.5125653297878,12.6822072133468 56.5449127274491,13.3407424424749 56.573784325367,14.0002231625192 56.5991709994144,14.660529681225 56.6210647008095,15.3215412627822 56.6394584686143,15.9831362768927 56.654346440595,16.6451923506331 56.6657238624055,17.307586522649 56.67358709506,17.9701953992046 56.677933620668,18.6328953115992 56.6787620464102,19.2955624744544 56.6760721067401,19.9580731443722 56.6698646638042,20.6203037784591 56.6601417060788,21.282131192215 56.6469063452276,21.943432716288 56.6301628111935,22.6040863516019 56.6099164455407,23.2639709223762 56.5861736930735,23.922966226566 56.5589420917603,24.5809531832687 56.5282302610022,25.2378139766594 56.4940478882858,25.8934321960358 56.4564057142701,26.5476929715805 56.4153155163602,27.2004831054654 56.3707900908252,27.8516911979552 56.3228432335229,28.5012077681911 56.2714897192993,29.1489253693643 56.2167452801302,29.7947386980206 56.1586265820819,30.4385446972665 56.0971512011646,31.0802426536785 56.0323375981587,31.7197342877491 55.9642050924945,32.3569238377352 55.8927738352675,32.9917181368037 55.8180647814723,33.6240266834038 55.74009966154,34.2537617048216 55.6589009522625,34.8808382139074 55.5744918471876,35.5051740589896 55.4868962265697,36.1266899670207 55.3961386269571,36.7453095800251 55.3022442104976,37.3609594849451 55.2052387340427,37.9735692370026 55.1051485181267,38.5830713767178 55.0020004158976,39.1894014407465 54.8958217820713,39.7924979667135 54.7866404419798,40.3923024922398 54.6744846607816,40.9887595483734 54.5593831128969,41.5818166476476 54.4413648517294,42.1714242670021 54.3204592797319,42.757535825811 54.1966961188706,43.3401076592708 54.0701053815371,43.919098987406 53.9407173419567,44.4944718799548 53.8085625081347,45.0661912174019 53.673671594382,45.6342246484243 53.5360754944551,46.198542544017 53.3958052553427,46.7591179485663 53.2528920517295,47.3159265281308 53.1073671611612,47.8689465161932 52.9592619399335,48.4181586571351 52.8086077997244,48.9635461476859 52.6554361849853,49.5050945765883 52.499778551104,50.0427918627159 52.3416663433486,50.5766281918714 52.1811309766006,51.1065959524853 52.0182038158815,51.6326896704254 51.8529161576737,52.1549059431199 51.6852992120372,52.673243373185 51.5153840855177,53.1877025017413 51.3432017648431,53.6982857415906 51.1687831014009,54.2049973104167 50.9921587964881,54.7078431641625 50.8133593873253,55.2068309307272 50.6324152338211,55.7019698441171 50.4493565060761,56.1932706791714 50.2642131726125,56.6807456869812 50.0770149893128,57.1644085311015 49.8877914890534,57.6442742246566 49.6965719720156,58.1203590684218 49.5033854966561,58.5926805899637 49.3082608713202,59.0612574839055 49.1112266464775,59.5261095533829 48.9123111075629,59.9872576527434 48.7115422684016,60.4447236315382 48.5089478652008,60.8985302798459 48.3045553510865,61.3487012749643 48.0983918911668,61.7952611294973 47.8904843581013,62.2382351408597 47.6808593281578,62.6776493422177 47.4695430777354,63.1135304548766 47.2565615803358,63.5459058421237 47.0419405039633,63.9748034645285 46.8257052089336,64.4002518367009 46.6078807460742,64.822279985501 46.3884918552974,65.2409174096934 46.1675629645276,65.6561940410346 45.9451181889661,66.0681402067793 45.7211813306754,66.476786593589 45.4957758784676,66.8821642128236 45.2689250080781,67.284304367196 45.0406515826125,67.6832386187654 44.8109781532476,68.0789987582454 44.5799269601738,68.4716167756018 44.3475199337644,68.8611248319107 44.1137786959568,69.2475552324516 43.878724561833,69.6309404010034 43.6423785413868,70.0113128553159 43.4047613414637,70.388705183725 43.1658933678633,70.7631500228809 42.925794727592,71.1346800365587 42.6844852312539,71.5033278955199 42.4419843955718,71.8691262583921 42.1983114460249,72.2321077535378 41.953485319597,72.5923049618788 41.7075246676227,72.9497504006463 41.4604478587259,73.3044765080245 41.2122729818388,73.6565156286596 40.963017849297,74.0059 40.7127)

test/CDB_JenksBinsTest.sql

@@ -5,9 +5,8 @@ WITH data AS (
                  15.01, 14.99,
                  20.1, 19.9]::numeric[] AS s
 )
--- expectation is: 1, 5, 10, 15, 20
--- TODO: fix cdb_jenksbins to match ^^
-SELECT round(unnest(CDB_JenksBins(s, 5))) FROM data;
+SELECT unnest(CDB_JenksBins(s, 5)) FROM data;
+
 
 WITH data_nulls AS (
     SELECT Array[0.99, 1.0, 1.01,
@@ -18,6 +17,20 @@ WITH data_nulls AS (
                  null, null,
                  20.1, 19.9]::numeric[] AS s
 )
--- expectation is: 1, 5, 10, 15, 20
--- TODO: fix cdb_jenksbins to match ^^
-SELECT round(unnest(CDB_JenksBins(s, 5))) FROM data_nulls;
+SELECT unnest(CDB_JenksBins(s, 5)) FROM data_nulls;
+
+
+WITH data_inverse AS (
+    SELECT Array[0.99, 1.0, 1.01,
+                 4.99, 5.01,
+                 10.01, 10.01,
+                 15.01, 14.99,
+                 20.1, 19.9]::numeric[] AS s
+)
+SELECT unnest(CDB_JenksBins(s, 5, 0, true)) FROM data_inverse;
+
+
+WITH data_small AS (
+    SELECT Array[0.99, 1.0, 10.01, 10.01, 10.01, 10.01]::numeric[] AS s
+)
+SELECT unnest(CDB_JenksBins(s, 4)) FROM data_small;

test/CDB_JenksBinsTest_expect

@@ -1,10 +1,18 @@
-1
-5
-10
-20
-20
-1
-5
-10
-20
-
+1.01
+5.01
+10.01
+15.01
+20.1
+1.01
+5.01
+10.01
+15.01
+20.1
+0.99
+4.99
+10.01
+14.99
+19.9
+0.99
+1.0
+10.01

test/CDB_OverviewsTest.sql

@@ -2,6 +2,7 @@ SET client_min_messages TO error;
 \set VERBOSITY default
 
 \i test/overviews/fixtures.sql
+vacuum ANALYZE; -- Make sure there are metrics for ST_EstimatedExtent
 
 SELECT _CDB_Aggregable_Attributes_Expression('base_bare_t'::regclass);
 SELECT _CDB_Aggregated_Attributes_Expression('base_bare_t'::regclass);

test/CDB_OverviewsTest_expect

@@ -6,6 +6,7 @@ INSERT 0 1114
 CREATE TABLE
 INSERT 0 5
 SELECT 1114
+VACUUM
 
 
 

test/extension/test.sh

@@ -418,6 +418,36 @@ function test_cdb_tablemetadatatouch_fully_qualifies_names() {
     sql postgres 'DROP TABLE touch_invalidations'
 }
 
+function test_cdb_tablemetadata_text() {
+
+    #create and touch tables
+    sql "CREATE TABLE touch_ex_a (id int);"
+    sql "CREATE TABLE touch_ex_b (id int);"
+    sql "CREATE TABLE touch_ex_c (id int);"
+    sql postgres "SELECT CDB_TableMetadataTouch('touch_ex_a');"
+    sql postgres "SELECT CDB_TableMetadataTouch('touch_ex_b');"
+    sql postgres "SELECT CDB_TableMetadataTouch('touch_ex_c');"
+
+    #ensure there is 1 record per table
+    QUERY="SELECT COUNT(1) FROM (SELECT 1 FROM cdb_tablemetadata_text "
+    QUERY+="GROUP BY tabname HAVING COUNT(1) > 1) s;"
+    sql postgres "$QUERY" should "0"
+
+    #ensure timestamps are distinct and properly ordered
+    QUERY="SELECT (SELECT updated_at FROM CDB_TableMetadata_Text WHERE tabname='public.touch_ex_a')"
+    QUERY+="    < (SELECT updated_at FROM CDB_TableMetadata_Text WHERE tabname='public.touch_ex_b');"
+    sql postgres "$QUERY" should "t"
+    QUERY="SELECT (SELECT updated_at FROM CDB_TableMetadata_Text WHERE tabname='public.touch_ex_b')"
+    QUERY+="    < (SELECT updated_at FROM CDB_TableMetadata_Text WHERE tabname='public.touch_ex_c');"
+    sql postgres "$QUERY" should "t"
+
+    #cleanup
+    sql "DROP TABLE touch_ex_a;"
+    sql "DROP TABLE touch_ex_b;"
+    sql "DROP TABLE touch_ex_c;"
+
+}
+
 function test_cdb_column_names() {
     sql cdb_testmember_1 'CREATE TABLE cdb_testmember_1.table_cnames(c int, a int, r int, t int, o int);'
     sql cdb_testmember_2 'CREATE TABLE cdb_testmember_2.table_cnames(d int, b int);'
@@ -543,7 +573,12 @@ END
     DATABASE=fdw_target sql postgres "SELECT cdb_tablemetadatatouch('test_fdw.foo'::regclass);"
     DATABASE=fdw_target sql postgres "SELECT cdb_tablemetadatatouch('test_fdw.foo2'::regclass);"
 
-    sql postgres "SELECT cartodb.CDB_Conf_SetConf('fdws', '{\"test_fdw\": {\"server\": {\"host\": \"localhost\", \"dbname\": \"fdw_target\"},
+    # Add PGPORT to conf if it is set
+    PORT_SPEC=""
+    if [[ "$PGPORT" != "" ]] ; then
+        PORT_SPEC=", \"port\": \"$PGPORT\""
+    fi
+    sql postgres "SELECT cartodb.CDB_Conf_SetConf('fdws', '{\"test_fdw\": {\"server\": {\"host\": \"localhost\", \"dbname\": \"fdw_target\" $PORT_SPEC },
                                            \"users\": {\"public\": {\"user\": \"fdw_user\", \"password\": \"foobarino\"}}}}')"
 
     sql postgres "SELECT cartodb._CDB_Setup_FDW('test_fdw')"