Merge pull request #92 from HPI-Information-Systems/bugfix/sigmod

Do not require column label for Sigmod format
HPI-Information-Systems · Mar 16, 2021 · 2b8a0fb · 2b8a0fb
2 parents 66d311e + c41866d
commit 2b8a0fb
Show file tree

Hide file tree

Showing 2 changed files with 18 additions and 1 deletion.
diff --git a/docs/basic_usage/experiments.md b/docs/basic_usage/experiments.md
@@ -95,6 +95,24 @@ The open-source matching solution Magellan is widely used in research. We suppor
 
 For more information, see [here](../../sigmod2021).
 
+```csv
+left_instance_id,right_instance_id,label
+http://store.com/42,http://otherstore.net/af82,1
+http://store.com/243,http://otherstore.net/cn82,0
+...
+```
+
+The label _(1=duplicate, 0=non-duplicate)_ is **optional** in Snowman. A missing label will consider the pair as duplicates.
+
+Therefore, the following list would be interpreted as duplicates:
+
+```csv
+left_instance_id,right_instance_id
+http://store.com/42,http://otherstore.net/af82
+http://store.com/243,http://otherstore.net/cn82
+...
+```
+
 ### Proprietary Formats
 
 We support a range of proprietary experiment formats. A list of those can be found here:

diff --git a/wrapper/src/api/providers/experiment/experimentProvider/file/formats/sigmod2021.ts b/wrapper/src/api/providers/experiment/experimentProvider/file/formats/sigmod2021.ts
@@ -7,7 +7,6 @@ export class Sigmod2021ExperimentInserter extends CSVInserter {
   protected readonly requiredColumns: string[] = [
     'left_instance_id',
     'right_instance_id',
-    'label',
   ];
 
   protected addRow(row: {