diff --git a/duckreg/estimators.py b/duckreg/estimators.py
index a4666ec..e70e52d 100644
--- a/duckreg/estimators.py
+++ b/duckreg/estimators.py
@@ -439,27 +439,26 @@ def __init__(
def prepare_data(self):
# create_cohort_and_ever_treated_columns
- self.cohort_query = f"""
- ALTER TABLE {self.table_name} ADD COLUMN cohort INTEGER;
- UPDATE {self.table_name} SET cohort = (
+ self.temp_table_query = f"""
+ CREATE TEMP TABLE temp_{self.table_name} AS
+ SELECT *, NULL::INTEGER AS cohort, NULL::INTEGER AS ever_treated
+ FROM {self.table_name};
+ UPDATE temp_{self.table_name} SET cohort = (
SELECT MIN({self.time_col})
FROM {self.table_name} AS p2
- WHERE p2.{self.unit_col} = {self.table_name}.{self.unit_col} AND p2.{self.treatment_col} = 1
+ WHERE p2.{self.unit_col} = temp_{self.table_name}.{self.unit_col} AND
+ p2.{self.treatment_col} = 1
);
+ UPDATE temp_{self.table_name} SET cohort = NULL WHERE cohort = 2147483647;
+ UPDATE temp_{self.table_name} SET ever_treated = CASE WHEN cohort IS NOT NULL THEN 1 ELSE 0 END;
"""
- self.conn.execute(self.cohort_query)
- self.ever_treated_query = f"""
- UPDATE {self.table_name} SET cohort = NULL WHERE cohort = 2147483647; -- Set to NULL if never treated
- ALTER TABLE {self.table_name} ADD COLUMN ever_treated INTEGER;
- UPDATE {self.table_name} SET ever_treated = CASE WHEN cohort IS NOT NULL THEN 1 ELSE 0 END;
- """
- self.conn.execute(self.ever_treated_query)
+ self.conn.execute(self.temp_table_query)
# retrieve_num_periods_and_cohorts
self.num_periods = self.conn.execute(
- f"SELECT MAX({self.time_col}) FROM {self.table_name}"
+ f"SELECT MAX({self.time_col}) FROM temp_{self.table_name}"
).fetchone()[0]
cohorts = self.conn.execute(
- f"SELECT DISTINCT cohort FROM {self.table_name} WHERE cohort IS NOT NULL"
+ f"SELECT DISTINCT cohort FROM temp_{self.table_name} WHERE cohort IS NOT NULL"
).fetchall()
self.cohorts = [row[0] for row in cohorts]
# generate_time_dummies
@@ -506,7 +505,7 @@ def prepare_data(self):
-- Treated group interacted with treatment time dummies
{self.treatment_dummies}
FROM
- {self.table_name} p;
+ temp_{self.table_name} p;
"""
self.conn.execute(self.design_matrix_query)
diff --git a/notebooks/event_study.ipynb b/notebooks/event_study.ipynb
index 4237b57..1fabbeb 100644
--- a/notebooks/event_study.ipynb
+++ b/notebooks/event_study.ipynb
@@ -16,7 +16,7 @@
"data": {
"text/html": [
"\n",
- " \n",
+ " \n",
" \n",
@@ -56,7 +56,7 @@
"data": {
"text/html": [
"\n",
- " \n",
+ " \n",
" \n",
@@ -100,7 +100,7 @@
"from duckreg.estimators import DuckMundlak, DuckMundlakEventStudy\n",
"import duckdb\n",
"import pyfixest as pf\n",
- "\n",
+ "from pyfixest.did.visualize import panelview\n",
"np.random.seed(42)"
]
},
@@ -206,7 +206,7 @@
" \"Y_it\": Y_it,\n",
" }\n",
" )\n",
- " return df\n"
+ " return df"
]
},
{
@@ -224,7 +224,7 @@
{
"data": {
"text/plain": [
- ""
+ ""
]
},
"execution_count": 4,
@@ -256,7 +256,7 @@
},
{
"cell_type": "code",
- "execution_count": 4,
+ "execution_count": 5,
"metadata": {},
"outputs": [
{
@@ -283,6 +283,42 @@
"create_duckdb_database(df, db_name)"
]
},
+ {
+ "cell_type": "code",
+ "execution_count": 6,
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ ""
+ ]
+ },
+ "execution_count": 6,
+ "metadata": {},
+ "output_type": "execute_result"
+ },
+ {
+ "data": {
+ "image/png": "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",
+ "text/plain": [
+ ""
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ }
+ ],
+ "source": [
+ "panelview(\n",
+ " df,\n",
+ " unit=\"unit_id\", time=\"time_id\", treat=\"W_it\",\n",
+ " collapse_to_cohort=True,\n",
+ " sort_by_timing=True,\n",
+ " title = \"Treatment Assignment\",\n",
+ ")"
+ ]
+ },
{
"cell_type": "markdown",
"metadata": {},
@@ -301,27 +337,15 @@
},
{
"cell_type": "code",
- "execution_count": 5,
+ "execution_count": 7,
"metadata": {},
"outputs": [
{
"name": "stderr",
"output_type": "stream",
"text": [
- "100%|██████████| 100/100 [00:45<00:00, 2.18it/s]\n"
+ "100%|██████████| 100/100 [00:47<00:00, 2.08it/s]\n"
]
- },
- {
- "data": {
- "text/plain": [
- "point_estimate 0.215242\n",
- "standard_error 0.000443\n",
- "Name: 1, dtype: float64"
- ]
- },
- "execution_count": 5,
- "metadata": {},
- "output_type": "execute_result"
}
],
"source": [
@@ -336,8 +360,28 @@
" n_bootstraps=100,\n",
" seed = 42\n",
")\n",
- "mundlak.fit()\n",
- "\n",
+ "mundlak.fit()"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 8,
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ "point_estimate 0.215242\n",
+ "standard_error 0.000443\n",
+ "Name: 1, dtype: float64"
+ ]
+ },
+ "execution_count": 8,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
"mundlak_results = mundlak.summary()\n",
"\n",
"restab = pd.DataFrame(\n",
@@ -349,7 +393,80 @@
},
{
"cell_type": "code",
- "execution_count": 6,
+ "execution_count": 12,
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/html": [
+ "\n",
+ "\n",
+ "
\n",
+ " \n",
+ " \n",
+ " | \n",
+ " Estimate | \n",
+ " Std. Error | \n",
+ " t value | \n",
+ " Pr(>|t|) | \n",
+ " 2.5% | \n",
+ " 97.5% | \n",
+ "
\n",
+ " \n",
+ " | Coefficient | \n",
+ " | \n",
+ " | \n",
+ " | \n",
+ " | \n",
+ " | \n",
+ " | \n",
+ "
\n",
+ " \n",
+ " \n",
+ " \n",
+ " | W_it | \n",
+ " 0.215242 | \n",
+ " 0.000631 | \n",
+ " 341.272582 | \n",
+ " 0.0 | \n",
+ " 0.214005 | \n",
+ " 0.216478 | \n",
+ "
\n",
+ " \n",
+ "
\n",
+ "