Changed the default time format from datetime64[s] to datetime64[ns].

caballerofelipe · caballerofelipe · commit 48c3250f4180 · 2024-06-13T16:09:21.000-05:00
Some testing was done and a bug was filed, HDF5Store doesn't save or loads `datetime64[s]` right so it was changed to [ns] which is the format Pandas chooses when transforming to datetime. See the link for the bug: pandas-dev/pandas#59004 .
diff --git a/lib/simple_portfolio_ledger/src/simple_portfolio_ledger.py b/lib/simple_portfolio_ledger/src/simple_portfolio_ledger.py
@@ -249,11 +249,11 @@ def ledger(
             pd
             # Join DataFrames
             .concat(dfs_toconcat, join='inner', axis=1)
-            # Convert dates to datetime64[s]
+            # Convert dates to appropriate format
             .astype(
                 {
-                    'date_execution': 'datetime64[s]',
-                    'date_order': 'datetime64[s]',
+                    'date_execution': 'datetime64[ns]',
+                    'date_order': 'datetime64[ns]',
                 }
             )
             # Try to pass colums where dtype is object to a type like int64 or float64

Original file line number	Diff line number	Diff line change
`@@ -249,11 +249,11 @@ def ledger(`
`249`	`249`	`pd`
`250`	`250`	`# Join DataFrames`
`251`	`251`	`.concat(dfs_toconcat, join='inner', axis=1)`
`252`		`- # Convert dates to datetime64[s]`
	`252`	`+ # Convert dates to appropriate format`
`253`	`253`	`.astype(`
`254`	`254`	`{`
`255`		`- 'date_execution': 'datetime64[s]',`
`256`		`- 'date_order': 'datetime64[s]',`
	`255`	`+ 'date_execution': 'datetime64[ns]',`
	`256`	`+ 'date_order': 'datetime64[ns]',`
`257`	`257`	`}`
`258`	`258`	`)`
`259`	`259`	`# Try to pass colums where dtype is object to a type like int64 or float64`