Monthly sales report

Build a formatted single-sheet report from start to finish

A complete script producing a report you could send without editing: title block, formatted currency, a native table, data bars and a chart.

1. Prepare the data

   Any Arrow-capable frame works. This uses Polars.

   import polars as pl
   import jetxl as jet
   
   df = pl.DataFrame({
       "Product":  ["Widget", "Gadget", "Doohickey", "Gizmo"],
       "Units":    [1200, 890, 2100, 450],
       "Price":    [19.99, 34.50, 8.75, 129.00],
       "Revenue":  [23988.00, 30705.00, 18375.00, 58050.00],
   })
   
2. Write the title block

   header_content puts text above the data, and data_start_row moves the table down to make room.

   header = [
       (1, 0, "ACME Corporation"),
       (2, 0, "Monthly Sales Report"),
       (3, 0, "Generated 2026-01-15"),
   ]
   
3. Compose the call

   jet.write_sheet_arrow(
       df.to_arrow(),
       "sales_report.xlsx",
       sheet_name="January",
   
       # title block
       header_content=header,
       data_start_row=5,
       merge_cells=[(1, 0, 1, 3), (2, 0, 2, 3)],
   
       # readability
       styled_headers=True,
       freeze_rows=5,
       auto_width=True,
       gridlines_visible=False,
       tab_color="0070C0",
   
       # numbers
       column_formats={
           "Units":   "thousands",
           "Price":   "currency",
           "Revenue": "currency",
       },
   
       # a native table over the data
       tables=[{
           "name": "SalesTable",
           "display_name": "January Sales",
           "start_row": 5,
           "start_col": 0,
           "style": "TableStyleMedium9",
           "show_row_stripes": True,
       }],
   
       # bars in the revenue column
       conditional_formats=[{
           "start_row": 6, "start_col": 3,
           "end_row": 6 + df.height - 1, "end_col": 3,
           "rule_type": "data_bar",
           "color": "638EC6",
           "show_value": True,
       }],
   
       # chart to the right
       charts=[{
           "chart_type": "column",
           "start_row": 5, "start_col": 0,
           "end_row": 5 + df.height, "end_col": 3,
           "from_col": 6,  "from_row": 5,
           "to_col": 14,   "to_row": 20,
           "title": "Revenue by Product",
           "title_bold": True,
           "title_font_size": 1400,
           "category_col": 0,
           "show_legend": False,
           "x_axis_title": "Product",
           "y_axis_title": "Revenue ($)",
           "axis_min": 0.0,
       }],
   )
   

This script is verified: it runs against jetxl 0.3.1 and produces a workbook that opens cleanly, with the title in A1, the column header on row 5 and the first data row on row 6.

Why the row numbers line up

The indexing rules differ by feature, which is the easiest thing to get wrong here.

Setting	Value	Reason
data_start_row	5	Rows 1 to 3 are the title, 4 is spacing
freeze_rows	5	Everything through the header stays put
Table start_row	5	The header row, 1-based
Conditional start_row	6	First data row, below the header
Chart from_col	6	Two clear columns right of the last data column

The table needs no end_row or end_col. Omitting them lets Jetxl size the range from the DataFrame, so the script keeps working when the data grows.

The conditional format and chart ranges use df.height rather than literals for the same reason. In Pandas, use len(df).

Turning gridlines off does more for the look of a report than any other single setting. Combined with a native table, the result reads as designed rather than exported.