updates to union, subset, and sort

c/CHANGELOG.rst

@@ -7,10 +7,21 @@
 - Add ``parents`` to the individual table to enable recording of pedigrees
   (:user:`ivan-krukov`, :user:`benjeffery`, :issue:`852`, :pr:`1125`, :pr:`866`, :pr:`1153`, :pr:`1177`).
 
+- Added a ``tsk_table_collection_canonicalse`` method, that allows checking for equality between
+  tables that are equivalent up to reordering (:user:`petrelharp`, :user:`mufernando`, :pr:`1108`).
+
+- Removed a previous requirement on ``tsk_table_collection_union``, allowing for unioning of
+  new information both above and below shared history (:user:`petrelharp`, :user:`mufernando`, :pr:`1108`).
+
 **Breaking changes**
 
 - Method ``tsk_individual_table_add_row`` has an extra arguments ``parents`` and ``parents_length``.
 
+**Breaking changes**
+
+- Add an ``options`` argument to ``tsk_table_collection_subset`` (:user:`petrelharp`, :pr:`1108`),
+  to allow for retaining the order of populations.
+
 **Bugfixes**
 
 ----------------------

c/tests/testlib.c

@@ -543,6 +543,7 @@ parse_mutations(const char *text, tsk_mutation_table_t *mutation_table)
     double time;
     char derived_state[MAX_LINE];
 
+    /* site, node, derived_state, [parent, time] */
     c = 0;
     while (text[c] != '\0') {
         /* Fill in the line */
@@ -638,30 +639,34 @@ parse_individuals(const char *text, tsk_individual_table_t *individual_table)
             q = strtok_r(NULL, ",", &q_cont);
         }
         CU_ASSERT_FATAL(q == NULL);
+
+        /* parents and name are optional */
         p = strtok_r(NULL, whitespace, &p_cont);
-        // the parents are comma-separated
-        parents_len = 1;
-        q = p;
-        while (*q != '\0') {
-            if (*q == ',') {
-                parents_len++;
+        parents_len = 0;
+        name = "";
+        if (p != NULL) {
+            // the parents are comma-separated
+            parents_len = 1;
+            q = p;
+            while (*q != '\0') {
+                if (*q == ',') {
+                    parents_len++;
+                }
+                q++;
+            }
+            CU_ASSERT_FATAL(parents_len >= 1);
+            strncpy(sub_line, p, MAX_LINE);
+            q = strtok_r(sub_line, ",", &q_cont);
+            for (k = 0; k < parents_len; k++) {
+                CU_ASSERT_FATAL(q != NULL);
+                parents[k] = atoi(q);
+                q = strtok_r(NULL, ",", &q_cont);
+            }
+            CU_ASSERT_FATAL(q == NULL);
+            p = strtok_r(NULL, whitespace, &p_cont);
+            if (p != NULL) {
+                name = p;
             }
-            q++;
-        }
-        CU_ASSERT_FATAL(parents_len >= 1);
-        strncpy(sub_line, p, MAX_LINE);
-        q = strtok_r(sub_line, ",", &q_cont);
-        for (k = 0; k < parents_len; k++) {
-            CU_ASSERT_FATAL(q != NULL);
-            parents[k] = atoi(q);
-            q = strtok_r(NULL, ",", &q_cont);
-        }
-        CU_ASSERT_FATAL(q == NULL);
-        p = strtok_r(NULL, whitespace, &p_cont);
-        if (p == NULL) {
-            name = "";
-        } else {
-            name = p;
         }
         ret = tsk_individual_table_add_row(individual_table, flags, location,
             location_len, parents, parents_len, name, strlen(name));

c/tskit/core.c

@@ -357,10 +357,6 @@ tsk_strerror_internal(int err)
         case TSK_ERR_NONBINARY_MUTATIONS_UNSUPPORTED:
             ret = "Only binary mutations are supported for this operation";
             break;
-        case TSK_ERR_UNION_NOT_SUPPORTED:
-            ret = "Union is not supported for cases where there is non-shared"
-                  "history older than the shared history of the two Table Collections";
-            break;
 
         /* Stats errors */
         case TSK_ERR_BAD_NUM_WINDOWS:

c/tskit/core.h

@@ -274,7 +274,6 @@ not found in the file.
 #define TSK_ERR_SORT_OFFSET_NOT_SUPPORTED                           -803
 #define TSK_ERR_NONBINARY_MUTATIONS_UNSUPPORTED                     -804
 #define TSK_ERR_MIGRATIONS_NOT_SUPPORTED                            -805
-#define TSK_ERR_UNION_NOT_SUPPORTED                                 -806
 
 /* Stats errors */
 #define TSK_ERR_BAD_NUM_WINDOWS                                     -900