Initial implementation of excel support (no tests)

README.md

@@ -15,6 +15,7 @@ OctoSQL is a query tool that allows you to join, analyse and transform data from
   - [PostgreSQL](#postgresql)
   - [MySQL](#mysql)
   - [Redis](#redis)
+  - [Excel](#excel)
 - [Documentation](#documentation)
 - [Architecture](#architecture)
 - [Datasource Pushdown Operations](#datasource-pushdown-operations)
@@ -123,6 +124,18 @@ CSV file seperated using commas. The first row should contain column names.
 ##### options:
 - path - path to file containing the data, required
 
+---
+#### Excel
+A single, coherent table in a Excel spreadsheet.
+The data may or may not have column names as first row.
+The data can be in any sheet, and start at any point, but it must
+be coherent (no spaces between rows, no spaces between columns).
+##### options:
+- path - path to file, required
+- columns - does the first row contain column names, optional: defaults to true
+- sheet - name of the sheet in which data is stored, optional: defaults to "Sheet1"
+- root - name of cell (i.e "A3", "BA14") which is the leftmost cell of the first
+row, optional: defaults to "A1"
 ---
 #### PostgreSQL
 Single PostgreSQL database table.

cmd/octosql/main.go

@@ -3,6 +3,7 @@ package main
 import (
 	"context"
 	"fmt"
+	"github.com/cube2222/octosql/storage/excel"
 	"log"
 	"os"
 	"reflect"
@@ -52,6 +53,7 @@ With OctoSQL you don't need O(n) client tools or a large data analysis system de
 				"mysql":    mysql.NewDataSourceBuilderFactoryFromConfig,
 				"postgres": postgres.NewDataSourceBuilderFactoryFromConfig,
 				"redis":    redis.NewDataSourceBuilderFactoryFromConfig,
+				"excel":    excel.NewDataSourceBuilderFactoryFromConfig,
 			},
 			cfg,
 		)

storage/excel/cell.go

@@ -0,0 +1,99 @@
+package excel
+
+import (
+	"github.com/pkg/errors"
+	"strconv"
+	"strings"
+	"unicode"
+)
+
+var ErroneousCell = newCell("_", 0)
+
+type cell struct {
+	column string
+	row    int
+}
+
+func newCell(col string, row int) cell {
+	return cell{
+		column: col,
+		row:    row,
+	}
+}
+
+func (cell cell) getCellName() string {
+	return cell.column + strconv.Itoa(cell.row)
+}
+
+func isAllZs(columnName string) bool {
+	for _, char := range columnName {
+		if char != 'Z' {
+			return false
+		}
+	}
+
+	return true
+}
+
+func (cell cell) getCellToTheRight() cell {
+	columnName := cell.column
+
+	if isAllZs(columnName) {
+		newColumnName := strings.Repeat("A", len(columnName)+1)
+		return newCell(newColumnName, cell.row)
+	}
+
+	index := len(columnName)
+
+	newColumnName := ""
+
+	for i := len(columnName) - 1; i >= 0; i-- {
+		index--
+
+		if columnName[i] == 'Z' {
+			newColumnName += "A"
+		} else {
+			newColumnName = string(columnName[i]+1) + newColumnName
+			break
+		}
+	}
+
+	newColumnName = columnName[:index] + newColumnName
+
+	return newCell(newColumnName, cell.row)
+}
+
+func (cell cell) getCellBelow() cell {
+	return newCell(cell.column, cell.row+1)
+}
+
+func isCorrectChar(char rune) bool {
+	return unicode.IsUpper(char) || unicode.IsDigit(char)
+}
+
+func getCellFromName(cell string) (cell, error) {
+
+	for index, char := range cell {
+		if !isCorrectChar(char) {
+			return ErroneousCell, errors.Errorf("The character %v can't make up for column name", char)
+		}
+
+		if unicode.IsDigit(char) {
+			if index == 0 {
+				return ErroneousCell, errors.Errorf("The first char of a cell can't be a digit")
+			}
+
+			columnName := cell[:index]
+			row := cell[index:]
+
+			rowNumeric, err := strconv.Atoi(row)
+			if err != nil {
+				return ErroneousCell, errors.Wrap(err, "couldn't parse row of cell")
+			}
+
+			return newCell(columnName, rowNumeric), nil
+		}
+	}
+
+	return ErroneousCell, errors.Errorf("cell %s has no row number", cell)
+}