diff --git a/fastetl/operators/db_to_csv_operator.py b/fastetl/operators/db_to_csv_operator.py
index 4547424..1516f9e 100644
--- a/fastetl/operators/db_to_csv_operator.py
+++ b/fastetl/operators/db_to_csv_operator.py
@@ -10,7 +10,10 @@
 from pathlib import Path
 from typing import Optional
 
+import pandas as pd
+
 from airflow.models.baseoperator import BaseOperator
+from airflow.utils.context import Context
 
 from fastetl.custom_functions.utils.get_table_cols_name import get_table_cols_name
 from fastetl.custom_functions.utils.db_connection import get_hook_and_engine_by_provider
@@ -84,7 +87,18 @@ def __init__(
         self.columns_to_remove: Optional[list[str]] = columns_to_remove
         self.int_columns: Optional[list[str]] = int_columns
 
-    def select_all_sql(self):
+    def _select_all_sql(self) -> str:
+        """Generate a SELECT statement to fetch all columns from
+        a table.
+
+        Returns:
+            str: a SELECT statement for all columns
+        """
+        if any(argument is None for argument in (self.table_scheme, self.table_name)):
+            raise ValueError(
+                "table_scheme and table_name are required "
+                "when select_sql is not provided."
+            )
         cols = get_table_cols_name(self.conn_id, self.table_scheme, self.table_name)
         if self.columns_to_remove:
             cols = [c for c in cols if c not in self.columns_to_remove]
@@ -95,18 +109,76 @@ def select_all_sql(self):
             FROM {self.table_scheme}.{self.table_name};
             """
 
-    def execute(self, context):
+    def _starts_with_sql_keyword(self, value: str) -> bool:
+        """Check if a string starts with a SQL keyword."""
+        sql_keywords = ("SELECT", "WITH", "INSERT", "UPDATE", "DELETE")
+        return any(value.strip().upper().startswith(kw) for kw in sql_keywords)
+
+    def _resolve_select_sql(self, select_sql: Optional[str | Path]) -> str:
+        """Resolve select_sql to a query string, handling files and inline
+        queries.
+
+        Args:
+            select_sql (Optional[str | Path]): The SQL query or file path to
+                resolve. If a string, it can be either a file path or an inline
+                SQL query. If a file path, the file must exist.
+
+        Raises:
+            FileNotFoundError: if a file path is specified but the file does not
+                exist.
+            TypeError: if select_sql is neither a string nor a Path object.
+            ValueError: if select_sql is None or empty and table_scheme or
+                table_name are not provided.
+
+        Returns:
+            str: the resolved SQL query string.
+        """
+
+        if not select_sql:
+            return self._select_all_sql()
+
+        # Handle Path objects
+        if isinstance(select_sql, Path):
+            if select_sql.is_file():
+                return select_sql.read_text(encoding="utf-8")
+            raise FileNotFoundError(f"File not found: {select_sql}")
+
+        # Handle strings
+        if isinstance(select_sql, str):
+            # Quick check: SQL keywords indicate it's a query, not a path
+            if self._starts_with_sql_keyword(select_sql):
+                return select_sql
+
+            # Attempt to load as a file path
+            try:
+                path = Path(select_sql)
+                if path.is_file():
+                    return path.read_text(encoding="utf-8")
+            except OSError:
+                # Catches "File name too long" and other path-related OS errors
+                # Treat as a query string instead
+                pass
+
+            # Default: treat as a query string
+            return select_sql
+
+        raise TypeError(f"select_sql must be a string or Path, got {type(select_sql)}")
+
+    def execute(self, context: Context):
+        """Executes the SQL query and saves the CSV file. In the process,
+        converts data types to integers and removes specified characters
+        if the options have been specified.
+
+        Args:
+            context (Context): The Airflow context object.
+
+        Returns:
+            str: The name of the written file.
+        """
+        _ = context  # left unused
         db_hook, _ = get_hook_and_engine_by_provider(self.conn_id)
 
-        if self.select_sql:
-            path_sql = Path(self.select_sql)
-            if path_sql.is_file():
-                query = path_sql.read_text(encoding="utf-8")
-            else:
-                query = self.select_sql
-            df_select = query
-        else:
-            df_select = self.select_all_sql()
+        df_select = self._resolve_select_sql(self.select_sql)
 
         self.log.info(f"Executing SQL check: {df_select}")
         df = db_hook.get_pandas_df(df_select)
@@ -114,7 +186,7 @@ def execute(self, context):
         # Convert columns data types to int
         if self.int_columns:
             for col in self.int_columns:
-                df[col] = df[col].astype("Int64")
+                df[col] = pd.to_numeric(df[col], errors="coerce").astype("Int64")
 
         # Remove specified characters
         if self.characters_to_remove: