Merge pull request PSLmodels#252 from PSLmodels/unique-RECID-values

Ensure RECID values are unique
donboyd5 · Oct 24, 2024 · f1dbf38 · f1dbf38
2 parents 903a89b + c31f6ad
commit f1dbf38
Show file tree

Hide file tree

Showing 2 changed files with 6 additions and 1 deletion.
diff --git a/Makefile b/Makefile
@@ -39,7 +39,9 @@ tmd/storage/output/tmd_weights.csv.gz: \
 tmd/storage/output/cached_files: \
   tmd/storage/output/tmd.csv.gz \
   tmd/storage/output/tmd_growfactors.csv \
-  tmd/storage/output/tmd_weights.csv.gz
+  tmd/storage/output/tmd_weights.csv.gz \
+  tmd/storage/__init__.py \
+  tmd/create_taxcalc_cached_files.py
 	python tmd/create_taxcalc_cached_files.py
 
 .PHONY=tmd_files

diff --git a/tmd/datasets/tmd.py b/tmd/datasets/tmd.py
@@ -29,6 +29,9 @@ def create_tmd_2021():
     print("Combining PUF and CPS nonfilers...")
     combined = pd.concat([tc_puf_21, tc_cps_21], ignore_index=True)
 
+    # ensure RECID values are unique
+    combined["RECID"] = np.arange(1, len(combined) + 1, dtype=int)
+
     trace1("A", combined)
 
     print("Adding Tax-Calculator outputs for 2021...")