Initial word table SQLite importer

2026-06-24 16:21:57 +08:00
commit 9fab4c73ed
7 changed files with 1074 additions and 0 deletions
--- a/.gitignore
+++ b/.gitignore
@@ -0,0 +1,6 @@
+.DS_Store
+bin/
+*.sqlite
+*.sqlite3
+*.db
+*.csv
--- a/HOWTO.md
+++ b/HOWTO.md
@@ -0,0 +1,201 @@
+# HOWTO: Word 表格导入 SQLite 并导出字段化 CSV
+
+这份说明面向实际使用：把报备表 `.docx` 导入 SQLite，然后把自动识别出的字段内容导出成 CSV，方便用 WPS/Excel 查看。
+
+## 1. 准备文件
+
+需要两个文件：
+
+- `wordtable2sqlite`：macOS 可执行文件。
+- `要导入的 .docx`：例如重要服务事项报备表。
+
+在 Windows 上使用时，对应可执行文件是：
+
+- `wordtable2sqlite-windows-amd64.exe`
+
+## 2. 导入单个 Word 文件
+
+macOS 示例：
+
+```bash
+./bin/wordtable2sqlite -reset -db report.sqlite input.docx
+```
+
+Windows 示例：
+
+```powershell
+.\wordtable2sqlite-windows-amd64.exe -reset -db report.sqlite input.docx
+```
+
+参数含义：
+
+- `-db report.sqlite`：指定输出的 SQLite 数据库文件。
+- `-reset`：导入前删除旧数据库，适合重新生成结果。
+- `input.docx`：要导入的 Word 文件路径。
+
+执行成功后会看到类似输出：
+
+```text
+imported input.docx: 1 tables, 31 cells, 15 fields
+database: report.sqlite
+```
+
+含义是：
+
+- `tables`：识别到几张 Word 表格。
+- `cells`：保存了多少个原始单元格。
+- `fields`：自动抽取了多少个字段化内容。
+
+## 3. 批量导入多个 Word 文件
+
+把多个 `.docx` 导入到同一个数据库：
+
+```bash
+./bin/wordtable2sqlite -reset -db reports.sqlite *.docx
+```
+
+如果是后续追加导入，不想清空旧数据，就不要加 `-reset`：
+
+```bash
+./bin/wordtable2sqlite -db reports.sqlite another.docx
+```
+
+## 4. 数据库里有什么表
+
+导入后会生成 4 张主要表：
+
+- `documents`：每个导入的 Word 文件。
+- `tables`：每个 Word 文件里的表格。
+- `cells`：原始单元格内容，保留行号、列号、合并单元格信息。
+- `fields`：字段化内容，最适合业务人员查看和检索。
+
+一般看业务内容，优先看 `fields`。
+
+## 5. 查看字段化内容
+
+在 macOS/Linux 终端中：
+
+```bash
+sqlite3 report.sqlite
+```
+
+进入 SQLite 后执行：
+
+```sql
+.headers on
+.mode line
+SELECT key, value FROM fields ORDER BY id;
+```
+
+`.mode line` 适合看长文本。每条记录会显示成：
+
+```text
+  key = 事件过程
+value = ...
+```
+
+如果想看表格模式：
+
+```sql
+.headers on
+.mode box
+SELECT key, value FROM fields ORDER BY id;
+```
+
+## 6. 导出字段化内容 CSV
+
+这是最常用的导出方式：
+
+```bash
+sqlite3 -header -csv report.sqlite "SELECT key, value FROM fields ORDER BY id;" > report-fields.csv
+```
+
+生成的 `report-fields.csv` 可以直接用 WPS/Excel 打开。
+
+如果打开后中文乱码，使用 WPS/Excel 的“从文本/CSV 导入”，编码选择：
+
+```text
+UTF-8
+```
+
+## 7. 导出原始单元格 CSV
+
+如果要检查 Word 表格的原始行列结构，导出 `cells`：
+
+```bash
+sqlite3 -header -csv report.sqlite "SELECT table_index, row_index, cell_index, col_index, col_span, text FROM cells ORDER BY table_index, row_index, cell_index;" > report-cells.csv
+```
+
+字段含义：
+
+- `table_index`：第几张表，从 0 开始。
+- `row_index`：第几行，从 0 开始。
+- `cell_index`：该行第几个单元格。
+- `col_index`：逻辑列号。
+- `col_span`：横向合并了几列。
+- `text`：单元格文字。
+
+## 8. 常用查询
+
+查所有字段名：
+
+```sql
+SELECT key FROM fields ORDER BY id;
+```
+
+查某个字段：
+
+```sql
+SELECT value FROM fields WHERE key = '工单编号';
+```
+
+按文件查看字段：
+
+```sql
+SELECT d.filename, f.key, f.value
+FROM fields f
+JOIN documents d ON d.id = f.document_id
+ORDER BY d.id, f.id;
+```
+
+查原始单元格：
+
+```sql
+SELECT row_index, cell_index, text
+FROM cells
+ORDER BY table_index, row_index, cell_index;
+```
+
+## 9. 推荐查看方式
+
+给业务人员看：
+
+1. 先导出 `report-fields.csv`。
+2. 用 WPS/Excel 打开。
+3. 每一行就是一个字段：`key` 是字段名，`value` 是字段内容。
+
+给开发或调试看：
+
+1. 打开 `report.sqlite`。
+2. 查看 `cells` 判断 Word 原始表格结构。
+3. 查看 `fields` 判断字段抽取是否符合预期。
+
+## 10. 本次样例命令
+
+本次样例数据库：
+
+```bash
+/Users/zhaoyilun/Documents/Codex/2026-06-24/wo-x/outputs/sample-baiyin.sqlite
+```
+
+导出字段化 CSV：
+
+```bash
+sqlite3 -header -csv /Users/zhaoyilun/Documents/Codex/2026-06-24/wo-x/outputs/sample-baiyin.sqlite "SELECT key, value FROM fields ORDER BY id;" > /Users/zhaoyilun/Documents/Codex/2026-06-24/wo-x/outputs/sample-baiyin-fields.csv
+```
+
+导出原始单元格 CSV：
+
+```bash
+sqlite3 -header -csv /Users/zhaoyilun/Documents/Codex/2026-06-24/wo-x/outputs/sample-baiyin.sqlite "SELECT table_index, row_index, cell_index, col_index, col_span, text FROM cells ORDER BY table_index, row_index, cell_index;" > /Users/zhaoyilun/Documents/Codex/2026-06-24/wo-x/outputs/sample-baiyin-cells.csv
+```
--- a/README.md
+++ b/README.md
@@ -0,0 +1,60 @@
+# wordtable2sqlite
+
+把 `.docx` 里的 Word 表格内容导入 SQLite 的最小 Go 工具。
+
+详细操作说明见：[HOWTO.md](HOWTO.md)
+
+## 特点
+
+- 不依赖 Word、WPS、Python 或 LibreOffice。
+- 直接解析 `.docx` 内部的 `word/document.xml`。
+- SQLite 使用纯 Go 驱动 `modernc.org/sqlite`，方便编译成单文件可执行程序。
+- 保底保存所有表格单元格到 `cells`。
+- 对常见报备表这种“标题行 + 表头行 + 数据行”的结构，自动生成 `fields`，便于按字段检索。
+
+## 编译
+
+```bash
+go build -o bin/wordtable2sqlite .
+```
+
+给非技术用户使用时，分发 `bin/wordtable2sqlite` 这个可执行文件即可。
+
+## 使用
+
+```bash
+./bin/wordtable2sqlite -reset -db data/report.sqlite input.docx
+```
+
+批量导入多个文件：
+
+```bash
+./bin/wordtable2sqlite -db data/report.sqlite *.docx
+```
+
+参数：
+
+- `-db`：输出 SQLite 文件路径，默认 `word_tables.sqlite`
+- `-reset`：导入前删除旧数据库
+- `-quiet`：不输出每个文件的摘要
+
+## 数据表
+
+- `documents`：每个导入的 Word 文件。
+- `tables`：Word 文档中的表格。
+- `cells`：每个表格单元格，包含表格序号、行号、列号、跨列数和文本。
+- `fields`：自动推断出的字段，常用于报备表检索。
+
+常用查询：
+
+```sql
+SELECT key, value
+FROM fields
+ORDER BY id;
+```
+
+```sql
+SELECT table_index, row_index, cell_index, text
+FROM cells
+ORDER BY table_index, row_index, cell_index;
+```
--- a/go.mod
+++ b/go.mod
@@ -0,0 +1,18 @@
+module wordtable2sqlite
+
+go 1.26
+
+require modernc.org/sqlite v1.40.1
+
+require (
+	github.com/dustin/go-humanize v1.0.1 // indirect
+	github.com/google/uuid v1.6.0 // indirect
+	github.com/mattn/go-isatty v0.0.20 // indirect
+	github.com/ncruces/go-strftime v0.1.9 // indirect
+	github.com/remyoudompheng/bigfft v0.0.0-20230129092748-24d4a6f8daec // indirect
+	golang.org/x/exp v0.0.0-20250620022241-b7579e27df2b // indirect
+	golang.org/x/sys v0.36.0 // indirect
+	modernc.org/libc v1.66.10 // indirect
+	modernc.org/mathutil v1.7.1 // indirect
+	modernc.org/memory v1.11.0 // indirect
+)
--- a/go.sum
+++ b/go.sum
@@ -0,0 +1,49 @@
+github.com/dustin/go-humanize v1.0.1 h1:GzkhY7T5VNhEkwH0PVJgjz+fX1rhBrR7pRT3mDkpeCY=
+github.com/dustin/go-humanize v1.0.1/go.mod h1:Mu1zIs6XwVuF/gI1OepvI0qD18qycQx+mFykh5fBlto=
+github.com/google/pprof v0.0.0-20250317173921-a4b03ec1a45e h1:ijClszYn+mADRFY17kjQEVQ1XRhq2/JR1M3sGqeJoxs=
+github.com/google/pprof v0.0.0-20250317173921-a4b03ec1a45e/go.mod h1:boTsfXsheKC2y+lKOCMpSfarhxDeIzfZG1jqGcPl3cA=
+github.com/google/uuid v1.6.0 h1:NIvaJDMOsjHA8n1jAhLSgzrAzy1Hgr+hNrb57e+94F0=
+github.com/google/uuid v1.6.0/go.mod h1:TIyPZe4MgqvfeYDBFedMoGGpEw/LqOeaOT+nhxU+yHo=
+github.com/mattn/go-isatty v0.0.20 h1:xfD0iDuEKnDkl03q4limB+vH+GxLEtL/jb4xVJSWWEY=
+github.com/mattn/go-isatty v0.0.20/go.mod h1:W+V8PltTTMOvKvAeJH7IuucS94S2C6jfK/D7dTCTo3Y=
+github.com/ncruces/go-strftime v0.1.9 h1:bY0MQC28UADQmHmaF5dgpLmImcShSi2kHU9XLdhx/f4=
+github.com/ncruces/go-strftime v0.1.9/go.mod h1:Fwc5htZGVVkseilnfgOVb9mKy6w1naJmn9CehxcKcls=
+github.com/remyoudompheng/bigfft v0.0.0-20230129092748-24d4a6f8daec h1:W09IVJc94icq4NjY3clb7Lk8O1qJ8BdBEF8z0ibU0rE=
+github.com/remyoudompheng/bigfft v0.0.0-20230129092748-24d4a6f8daec/go.mod h1:qqbHyh8v60DhA7CoWK5oRCqLrMHRGoxYCSS9EjAz6Eo=
+golang.org/x/exp v0.0.0-20250620022241-b7579e27df2b h1:M2rDM6z3Fhozi9O7NWsxAkg/yqS/lQJ6PmkyIV3YP+o=
+golang.org/x/exp v0.0.0-20250620022241-b7579e27df2b/go.mod h1:3//PLf8L/X+8b4vuAfHzxeRUl04Adcb341+IGKfnqS8=
+golang.org/x/mod v0.27.0 h1:kb+q2PyFnEADO2IEF935ehFUXlWiNjJWtRNgBLSfbxQ=
+golang.org/x/mod v0.27.0/go.mod h1:rWI627Fq0DEoudcK+MBkNkCe0EetEaDSwJJkCcjpazc=
+golang.org/x/sync v0.16.0 h1:ycBJEhp9p4vXvUZNszeOq0kGTPghopOL8q0fq3vstxw=
+golang.org/x/sync v0.16.0/go.mod h1:1dzgHSNfp02xaA81J2MS99Qcpr2w7fw1gpm99rleRqA=
+golang.org/x/sys v0.6.0/go.mod h1:oPkhp1MJrh7nUepCBck5+mAzfO9JrbApNNgaTdGDITg=
+golang.org/x/sys v0.36.0 h1:KVRy2GtZBrk1cBYA7MKu5bEZFxQk4NIDV6RLVcC8o0k=
+golang.org/x/sys v0.36.0/go.mod h1:OgkHotnGiDImocRcuBABYBEXf8A9a87e/uXjp9XT3ks=
+golang.org/x/tools v0.36.0 h1:kWS0uv/zsvHEle1LbV5LE8QujrxB3wfQyxHfhOk0Qkg=
+golang.org/x/tools v0.36.0/go.mod h1:WBDiHKJK8YgLHlcQPYQzNCkUxUypCaa5ZegCVutKm+s=
+modernc.org/cc/v4 v4.26.5 h1:xM3bX7Mve6G8K8b+T11ReenJOT+BmVqQj0FY5T4+5Y4=
+modernc.org/cc/v4 v4.26.5/go.mod h1:uVtb5OGqUKpoLWhqwNQo/8LwvoiEBLvZXIQ/SmO6mL0=
+modernc.org/ccgo/v4 v4.28.1 h1:wPKYn5EC/mYTqBO373jKjvX2n+3+aK7+sICCv4Fjy1A=
+modernc.org/ccgo/v4 v4.28.1/go.mod h1:uD+4RnfrVgE6ec9NGguUNdhqzNIeeomeXf6CL0GTE5Q=
+modernc.org/fileutil v1.3.40 h1:ZGMswMNc9JOCrcrakF1HrvmergNLAmxOPjizirpfqBA=
+modernc.org/fileutil v1.3.40/go.mod h1:HxmghZSZVAz/LXcMNwZPA/DRrQZEVP9VX0V4LQGQFOc=
+modernc.org/gc/v2 v2.6.5 h1:nyqdV8q46KvTpZlsw66kWqwXRHdjIlJOhG6kxiV/9xI=
+modernc.org/gc/v2 v2.6.5/go.mod h1:YgIahr1ypgfe7chRuJi2gD7DBQiKSLMPgBQe9oIiito=
+modernc.org/goabi0 v0.2.0 h1:HvEowk7LxcPd0eq6mVOAEMai46V+i7Jrj13t4AzuNks=
+modernc.org/goabi0 v0.2.0/go.mod h1:CEFRnnJhKvWT1c1JTI3Avm+tgOWbkOu5oPA8eH8LnMI=
+modernc.org/libc v1.66.10 h1:yZkb3YeLx4oynyR+iUsXsybsX4Ubx7MQlSYEw4yj59A=
+modernc.org/libc v1.66.10/go.mod h1:8vGSEwvoUoltr4dlywvHqjtAqHBaw0j1jI7iFBTAr2I=
+modernc.org/mathutil v1.7.1 h1:GCZVGXdaN8gTqB1Mf/usp1Y/hSqgI2vAGGP4jZMCxOU=
+modernc.org/mathutil v1.7.1/go.mod h1:4p5IwJITfppl0G4sUEDtCr4DthTaT47/N3aT6MhfgJg=
+modernc.org/memory v1.11.0 h1:o4QC8aMQzmcwCK3t3Ux/ZHmwFPzE6hf2Y5LbkRs+hbI=
+modernc.org/memory v1.11.0/go.mod h1:/JP4VbVC+K5sU2wZi9bHoq2MAkCnrt2r98UGeSK7Mjw=
+modernc.org/opt v0.1.4 h1:2kNGMRiUjrp4LcaPuLY2PzUfqM/w9N23quVwhKt5Qm8=
+modernc.org/opt v0.1.4/go.mod h1:03fq9lsNfvkYSfxrfUhZCWPk1lm4cq4N+Bh//bEtgns=
+modernc.org/sortutil v1.2.1 h1:+xyoGf15mM3NMlPDnFqrteY07klSFxLElE2PVuWIJ7w=
+modernc.org/sortutil v1.2.1/go.mod h1:7ZI3a3REbai7gzCLcotuw9AC4VZVpYMjDzETGsSMqJE=
+modernc.org/sqlite v1.40.1 h1:VfuXcxcUWWKRBuP8+BR9L7VnmusMgBNNnBYGEe9w/iY=
+modernc.org/sqlite v1.40.1/go.mod h1:9fjQZ0mB1LLP0GYrp39oOJXx/I2sxEnZtzCmEQIKvGE=
+modernc.org/strutil v1.2.1 h1:UneZBkQA+DX2Rp35KcM69cSsNES9ly8mQWD71HKlOA0=
+modernc.org/strutil v1.2.1/go.mod h1:EHkiggD70koQxjVdSBM3JKM7k6L0FbGE5eymy9i3B9A=
+modernc.org/token v1.1.0 h1:Xl7Ap9dKaEs5kLoOQeQmPWevfnk/DM5qcLcYlA8ys6Y=
+modernc.org/token v1.1.0/go.mod h1:UGzOrNV1mAFSEB63lOFHIpNRUVMvYTc6yu1SMY/XTDM=
--- a/main.go
+++ b/main.go
@@ -0,0 +1,606 @@
+package main
+
+import (
+	"archive/zip"
+	"bytes"
+	"context"
+	"crypto/sha256"
+	"database/sql"
+	"encoding/hex"
+	"encoding/xml"
+	"errors"
+	"flag"
+	"fmt"
+	"io"
+	"os"
+	"path/filepath"
+	"strconv"
+	"strings"
+	"time"
+
+	_ "modernc.org/sqlite"
+)
+
+type Table struct {
+	Index int
+	Rows  []Row
+}
+
+type Row struct {
+	Index int
+	Cells []Cell
+}
+
+type Cell struct {
+	RowIndex  int
+	CellIndex int
+	ColIndex  int
+	ColSpan   int
+	VMerge    string
+	Text      string
+}
+
+type Field struct {
+	RecordIndex int
+	RowIndex    int
+	ColIndex    int
+	Key         string
+	Value       string
+	SourceKind  string
+}
+
+func main() {
+	if err := run(os.Args[1:]); err != nil {
+		fmt.Fprintln(os.Stderr, "error:", err)
+		os.Exit(1)
+	}
+}
+
+func run(args []string) error {
+	fs := flag.NewFlagSet("wordtable2sqlite", flag.ContinueOnError)
+	dbPath := fs.String("db", "word_tables.sqlite", "SQLite database path")
+	reset := fs.Bool("reset", false, "remove the target database before importing")
+	quiet := fs.Bool("quiet", false, "suppress per-file summary output")
+	if err := fs.Parse(args); err != nil {
+		return err
+	}
+	docxPaths := fs.Args()
+	if len(docxPaths) == 0 {
+		return errors.New("usage: wordtable2sqlite -db output.sqlite [-reset] file1.docx [file2.docx ...]")
+	}
+
+	if *reset {
+		if err := os.Remove(*dbPath); err != nil && !errors.Is(err, os.ErrNotExist) {
+			return fmt.Errorf("reset database: %w", err)
+		}
+	}
+	if dir := filepath.Dir(*dbPath); dir != "." {
+		if err := os.MkdirAll(dir, 0o755); err != nil {
+			return fmt.Errorf("create database directory: %w", err)
+		}
+	}
+
+	db, err := sql.Open("sqlite", *dbPath)
+	if err != nil {
+		return fmt.Errorf("open sqlite database: %w", err)
+	}
+	defer db.Close()
+
+	ctx := context.Background()
+	for _, docxPath := range docxPaths {
+		before, err := dbStats(ctx, db)
+		if err != nil {
+			return err
+		}
+		if err := importDocument(ctx, db, docxPath); err != nil {
+			return err
+		}
+		after, err := dbStats(ctx, db)
+		if err != nil {
+			return err
+		}
+		if !*quiet {
+			fmt.Printf("imported %s: %d tables, %d cells, %d fields\n",
+				docxPath,
+				after.Tables-before.Tables,
+				after.Cells-before.Cells,
+				after.Fields-before.Fields,
+			)
+		}
+	}
+	if !*quiet {
+		fmt.Printf("database: %s\n", *dbPath)
+	}
+	return nil
+}
+
+type stats struct {
+	Tables int
+	Cells  int
+	Fields int
+}
+
+func dbStats(ctx context.Context, db *sql.DB) (stats, error) {
+	if err := ensureSchema(ctx, db); err != nil {
+		return stats{}, err
+	}
+	var s stats
+	for _, item := range []struct {
+		query string
+		dst   *int
+	}{
+		{`SELECT COUNT(*) FROM tables`, &s.Tables},
+		{`SELECT COUNT(*) FROM cells`, &s.Cells},
+		{`SELECT COUNT(*) FROM fields`, &s.Fields},
+	} {
+		if err := db.QueryRowContext(ctx, item.query).Scan(item.dst); err != nil {
+			return stats{}, err
+		}
+	}
+	return s, nil
+}
+
+func importDocument(ctx context.Context, db *sql.DB, docxPath string) error {
+	if err := ensureSchema(ctx, db); err != nil {
+		return err
+	}
+	tables, err := readDocxTables(docxPath)
+	if err != nil {
+		return err
+	}
+	sum, err := fileSHA256(docxPath)
+	if err != nil {
+		return err
+	}
+
+	tx, err := db.BeginTx(ctx, nil)
+	if err != nil {
+		return err
+	}
+	defer tx.Rollback()
+
+	res, err := tx.ExecContext(ctx,
+		`INSERT INTO documents(path, filename, sha256, imported_at) VALUES (?, ?, ?, ?)`,
+		docxPath,
+		filepath.Base(docxPath),
+		sum,
+		time.Now().Format(time.RFC3339),
+	)
+	if err != nil {
+		return fmt.Errorf("insert document: %w", err)
+	}
+	documentID, err := res.LastInsertId()
+	if err != nil {
+		return err
+	}
+
+	for tableIndex, table := range tables {
+		cellCount := 0
+		for _, row := range table.Rows {
+			cellCount += len(row.Cells)
+		}
+		res, err := tx.ExecContext(ctx,
+			`INSERT INTO tables(document_id, table_index, row_count, cell_count) VALUES (?, ?, ?, ?)`,
+			documentID,
+			tableIndex,
+			len(table.Rows),
+			cellCount,
+		)
+		if err != nil {
+			return fmt.Errorf("insert table %d: %w", tableIndex, err)
+		}
+		tableID, err := res.LastInsertId()
+		if err != nil {
+			return err
+		}
+
+		for rowIndex, row := range table.Rows {
+			for cellIndex, cell := range row.Cells {
+				if _, err := tx.ExecContext(ctx,
+					`INSERT INTO cells(document_id, table_id, table_index, row_index, cell_index, col_index, col_span, v_merge, text)
+					 VALUES (?, ?, ?, ?, ?, ?, ?, ?, ?)`,
+					documentID,
+					tableID,
+					tableIndex,
+					rowIndex,
+					cellIndex,
+					cell.ColIndex,
+					cell.ColSpan,
+					cell.VMerge,
+					cell.Text,
+				); err != nil {
+					return fmt.Errorf("insert cell table=%d row=%d cell=%d: %w", tableIndex, rowIndex, cellIndex, err)
+				}
+			}
+		}
+
+		for _, field := range inferFields(table) {
+			if _, err := tx.ExecContext(ctx,
+				`INSERT INTO fields(document_id, table_id, table_index, record_index, row_index, col_index, key, value, source_kind)
+				 VALUES (?, ?, ?, ?, ?, ?, ?, ?, ?)`,
+				documentID,
+				tableID,
+				tableIndex,
+				field.RecordIndex,
+				field.RowIndex,
+				field.ColIndex,
+				field.Key,
+				field.Value,
+				field.SourceKind,
+			); err != nil {
+				return fmt.Errorf("insert field %q: %w", field.Key, err)
+			}
+		}
+	}
+
+	return tx.Commit()
+}
+
+func ensureSchema(ctx context.Context, db execer) error {
+	statements := []string{
+		`PRAGMA foreign_keys = ON`,
+		`CREATE TABLE IF NOT EXISTS documents (
+			id INTEGER PRIMARY KEY AUTOINCREMENT,
+			path TEXT NOT NULL,
+			filename TEXT NOT NULL,
+			sha256 TEXT NOT NULL,
+			imported_at TEXT NOT NULL
+		)`,
+		`CREATE TABLE IF NOT EXISTS tables (
+			id INTEGER PRIMARY KEY AUTOINCREMENT,
+			document_id INTEGER NOT NULL REFERENCES documents(id) ON DELETE CASCADE,
+			table_index INTEGER NOT NULL,
+			row_count INTEGER NOT NULL,
+			cell_count INTEGER NOT NULL
+		)`,
+		`CREATE TABLE IF NOT EXISTS cells (
+			id INTEGER PRIMARY KEY AUTOINCREMENT,
+			document_id INTEGER NOT NULL REFERENCES documents(id) ON DELETE CASCADE,
+			table_id INTEGER NOT NULL REFERENCES tables(id) ON DELETE CASCADE,
+			table_index INTEGER NOT NULL,
+			row_index INTEGER NOT NULL,
+			cell_index INTEGER NOT NULL,
+			col_index INTEGER NOT NULL,
+			col_span INTEGER NOT NULL,
+			v_merge TEXT NOT NULL DEFAULT '',
+			text TEXT NOT NULL
+		)`,
+		`CREATE TABLE IF NOT EXISTS fields (
+			id INTEGER PRIMARY KEY AUTOINCREMENT,
+			document_id INTEGER NOT NULL REFERENCES documents(id) ON DELETE CASCADE,
+			table_id INTEGER NOT NULL REFERENCES tables(id) ON DELETE CASCADE,
+			table_index INTEGER NOT NULL,
+			record_index INTEGER NOT NULL,
+			row_index INTEGER NOT NULL,
+			col_index INTEGER NOT NULL,
+			key TEXT NOT NULL,
+			value TEXT NOT NULL,
+			source_kind TEXT NOT NULL
+		)`,
+		`CREATE INDEX IF NOT EXISTS idx_cells_doc_table ON cells(document_id, table_index, row_index, cell_index)`,
+		`CREATE INDEX IF NOT EXISTS idx_fields_doc_key ON fields(document_id, key)`,
+	}
+	for _, statement := range statements {
+		if _, err := db.ExecContext(ctx, statement); err != nil {
+			return fmt.Errorf("schema statement failed: %w", err)
+		}
+	}
+	return nil
+}
+
+type execer interface {
+	ExecContext(context.Context, string, ...any) (sql.Result, error)
+}
+
+func readDocxTables(path string) ([]Table, error) {
+	zr, err := zip.OpenReader(path)
+	if err != nil {
+		return nil, fmt.Errorf("open docx zip: %w", err)
+	}
+	defer zr.Close()
+
+	for _, file := range zr.File {
+		if file.Name != "word/document.xml" {
+			continue
+		}
+		rc, err := file.Open()
+		if err != nil {
+			return nil, err
+		}
+		defer rc.Close()
+		data, err := io.ReadAll(rc)
+		if err != nil {
+			return nil, err
+		}
+		return parseDocumentXML(data)
+	}
+	return nil, errors.New("word/document.xml not found in docx")
+}
+
+func parseDocumentXML(data []byte) ([]Table, error) {
+	decoder := xml.NewDecoder(bytes.NewReader(data))
+	var tables []Table
+	for {
+		token, err := decoder.Token()
+		if errors.Is(err, io.EOF) {
+			return tables, nil
+		}
+		if err != nil {
+			return nil, err
+		}
+		start, ok := token.(xml.StartElement)
+		if !ok || start.Name.Local != "tbl" {
+			continue
+		}
+		table, err := parseTable(decoder)
+		if err != nil {
+			return nil, err
+		}
+		table.Index = len(tables)
+		tables = append(tables, table)
+	}
+}
+
+func parseTable(decoder *xml.Decoder) (Table, error) {
+	var table Table
+	for {
+		token, err := decoder.Token()
+		if err != nil {
+			return table, err
+		}
+		switch t := token.(type) {
+		case xml.StartElement:
+			if t.Name.Local == "tr" {
+				row, err := parseRow(decoder, len(table.Rows))
+				if err != nil {
+					return table, err
+				}
+				table.Rows = append(table.Rows, row)
+			}
+		case xml.EndElement:
+			if t.Name.Local == "tbl" {
+				return table, nil
+			}
+		}
+	}
+}
+
+func parseRow(decoder *xml.Decoder, rowIndex int) (Row, error) {
+	row := Row{Index: rowIndex}
+	logicalCol := 0
+	for {
+		token, err := decoder.Token()
+		if err != nil {
+			return row, err
+		}
+		switch t := token.(type) {
+		case xml.StartElement:
+			if t.Name.Local == "tc" {
+				cell, err := parseCell(decoder)
+				if err != nil {
+					return row, err
+				}
+				cell.RowIndex = rowIndex
+				cell.CellIndex = len(row.Cells)
+				cell.ColIndex = logicalCol
+				if cell.ColSpan < 1 {
+					cell.ColSpan = 1
+				}
+				logicalCol += cell.ColSpan
+				row.Cells = append(row.Cells, cell)
+			}
+		case xml.EndElement:
+			if t.Name.Local == "tr" {
+				return row, nil
+			}
+		}
+	}
+}
+
+func parseCell(decoder *xml.Decoder) (Cell, error) {
+	cell := Cell{ColSpan: 1}
+	var b strings.Builder
+	inText := false
+	tcDepth := 1
+	for {
+		token, err := decoder.Token()
+		if err != nil {
+			return cell, err
+		}
+		switch t := token.(type) {
+		case xml.StartElement:
+			switch t.Name.Local {
+			case "tc":
+				tcDepth++
+			case "gridSpan":
+				if tcDepth == 1 {
+					if span, err := strconv.Atoi(attr(t, "val")); err == nil && span > 0 {
+						cell.ColSpan = span
+					}
+				}
+			case "vMerge":
+				if tcDepth == 1 {
+					cell.VMerge = attr(t, "val")
+					if cell.VMerge == "" {
+						cell.VMerge = "continue"
+					}
+				}
+			case "t":
+				inText = true
+			case "tab":
+				b.WriteByte('\t')
+			case "br", "cr":
+				b.WriteByte('\n')
+			}
+		case xml.CharData:
+			if inText {
+				b.Write([]byte(t))
+			}
+		case xml.EndElement:
+			switch t.Name.Local {
+			case "t":
+				inText = false
+			case "p":
+				if b.Len() > 0 && !strings.HasSuffix(b.String(), "\n") {
+					b.WriteByte('\n')
+				}
+			case "tc":
+				tcDepth--
+				if tcDepth == 0 {
+					cell.Text = normalizeCellText(b.String())
+					return cell, nil
+				}
+			}
+		}
+	}
+}
+
+func attr(start xml.StartElement, local string) string {
+	for _, a := range start.Attr {
+		if a.Name.Local == local {
+			return a.Value
+		}
+	}
+	return ""
+}
+
+func normalizeCellText(s string) string {
+	s = strings.ReplaceAll(s, "\r\n", "\n")
+	s = strings.ReplaceAll(s, "\r", "\n")
+	lines := strings.Split(s, "\n")
+	out := make([]string, 0, len(lines))
+	for _, line := range lines {
+		line = strings.TrimSpace(strings.Join(strings.Fields(line), " "))
+		if line != "" {
+			out = append(out, line)
+		}
+	}
+	return strings.Join(out, "\n")
+}
+
+func inferFields(table Table) []Field {
+	headerIndex := detectHeaderRow(table)
+	if headerIndex >= 0 {
+		return inferHeaderFields(table, headerIndex)
+	}
+	return inferKeyValueFields(table)
+}
+
+func detectHeaderRow(table Table) int {
+	for i := 0; i < len(table.Rows)-1; i++ {
+		row := table.Rows[i]
+		if isTitleRow(row) {
+			continue
+		}
+		if nonEmptyCellCount(row) >= 2 && nonEmptyCellCount(table.Rows[i+1]) >= 2 {
+			return i
+		}
+	}
+	return -1
+}
+
+func isTitleRow(row Row) bool {
+	nonEmpty := nonEmptyCellCount(row)
+	if nonEmpty != 1 || len(row.Cells) == 0 {
+		return false
+	}
+	return len(row.Cells) == 1 || row.Cells[0].ColSpan > 1
+}
+
+func nonEmptyCellCount(row Row) int {
+	count := 0
+	for _, cell := range row.Cells {
+		if strings.TrimSpace(cell.Text) != "" {
+			count++
+		}
+	}
+	return count
+}
+
+func inferHeaderFields(table Table, headerIndex int) []Field {
+	headers := table.Rows[headerIndex].Cells
+	var fields []Field
+	recordIndex := 0
+	for rowIndex := headerIndex + 1; rowIndex < len(table.Rows); rowIndex++ {
+		row := table.Rows[rowIndex]
+		if nonEmptyCellCount(row) == 0 {
+			continue
+		}
+		for i, header := range headers {
+			if i >= len(row.Cells) {
+				continue
+			}
+			key := cleanFieldKey(header.Text)
+			if key == "" {
+				continue
+			}
+			value := strings.TrimSpace(row.Cells[i].Text)
+			fields = append(fields, Field{
+				RecordIndex: recordIndex,
+				RowIndex:    rowIndex,
+				ColIndex:    row.Cells[i].ColIndex,
+				Key:         key,
+				Value:       value,
+				SourceKind:  "header_row",
+			})
+		}
+		recordIndex++
+	}
+	return fields
+}
+
+func inferKeyValueFields(table Table) []Field {
+	var fields []Field
+	recordIndex := 0
+	for rowIndex, row := range table.Rows {
+		cells := nonEmptyCells(row)
+		if len(cells) < 2 || len(cells)%2 != 0 {
+			continue
+		}
+		for i := 0; i < len(cells); i += 2 {
+			key := cleanFieldKey(cells[i].Text)
+			if key == "" {
+				continue
+			}
+			fields = append(fields, Field{
+				RecordIndex: recordIndex,
+				RowIndex:    rowIndex,
+				ColIndex:    cells[i+1].ColIndex,
+				Key:         key,
+				Value:       cells[i+1].Text,
+				SourceKind:  "key_value",
+			})
+		}
+		recordIndex++
+	}
+	return fields
+}
+
+func nonEmptyCells(row Row) []Cell {
+	out := make([]Cell, 0, len(row.Cells))
+	for _, cell := range row.Cells {
+		if strings.TrimSpace(cell.Text) != "" {
+			out = append(out, cell)
+		}
+	}
+	return out
+}
+
+func cleanFieldKey(s string) string {
+	s = strings.TrimSpace(s)
+	s = strings.TrimRight(s, "：: \t\r\n")
+	return strings.TrimSpace(s)
+}
+
+func fileSHA256(path string) (string, error) {
+	f, err := os.Open(path)
+	if err != nil {
+		return "", err
+	}
+	defer f.Close()
+	h := sha256.New()
+	if _, err := io.Copy(h, f); err != nil {
+		return "", err
+	}
+	return hex.EncodeToString(h.Sum(nil)), nil
+}
--- a/main_test.go
+++ b/main_test.go
@@ -0,0 +1,134 @@
+package main
+
+import (
+	"archive/zip"
+	"context"
+	"database/sql"
+	"os"
+	"path/filepath"
+	"testing"
+
+	_ "modernc.org/sqlite"
+)
+
+func TestParseDocumentXMLExtractsTableCells(t *testing.T) {
+	xml := `<?xml version="1.0" encoding="UTF-8"?>
+<w:document xmlns:w="http://schemas.openxmlformats.org/wordprocessingml/2006/main">
+  <w:body>
+    <w:tbl>
+      <w:tr>
+        <w:tc><w:tcPr><w:gridSpan w:val="2"/></w:tcPr><w:p><w:r><w:t>报备表</w:t></w:r></w:p></w:tc>
+      </w:tr>
+      <w:tr>
+        <w:tc><w:p><w:r><w:t>序号</w:t></w:r></w:p></w:tc>
+        <w:tc><w:p><w:r><w:t>事件过程</w:t></w:r></w:p></w:tc>
+      </w:tr>
+      <w:tr>
+        <w:tc><w:p><w:r><w:t>1</w:t></w:r></w:p></w:tc>
+        <w:tc><w:p><w:r><w:t>起因</w:t><w:br/><w:t>处理经过</w:t></w:r></w:p></w:tc>
+      </w:tr>
+    </w:tbl>
+  </w:body>
+</w:document>`
+
+	tables, err := parseDocumentXML([]byte(xml))
+	if err != nil {
+		t.Fatalf("parseDocumentXML returned error: %v", err)
+	}
+	if len(tables) != 1 {
+		t.Fatalf("got %d tables, want 1", len(tables))
+	}
+	if got := tables[0].Rows[0].Cells[0].Text; got != "报备表" {
+		t.Fatalf("title cell text = %q, want 报备表", got)
+	}
+	if got := tables[0].Rows[0].Cells[0].ColSpan; got != 2 {
+		t.Fatalf("title cell colspan = %d, want 2", got)
+	}
+	if got := tables[0].Rows[2].Cells[1].Text; got != "起因\n处理经过" {
+		t.Fatalf("multiline text = %q, want 起因\\n处理经过", got)
+	}
+}
+
+func TestInferHeaderFieldsSkipsTitleRow(t *testing.T) {
+	table := Table{
+		Rows: []Row{
+			{Cells: []Cell{{Text: "报备表", ColSpan: 2}}},
+			{Cells: []Cell{{Text: "序号", ColSpan: 1}, {Text: "事件过程", ColSpan: 1}}},
+			{Cells: []Cell{{Text: "1", ColSpan: 1}, {Text: "客户反映设备位置不合理", ColSpan: 1}}},
+		},
+	}
+
+	fields := inferFields(table)
+	if len(fields) != 2 {
+		t.Fatalf("got %d fields, want 2", len(fields))
+	}
+	if fields[0].Key != "序号" || fields[0].Value != "1" || fields[0].SourceKind != "header_row" {
+		t.Fatalf("unexpected first field: %#v", fields[0])
+	}
+	if fields[1].Key != "事件过程" || fields[1].Value != "客户反映设备位置不合理" {
+		t.Fatalf("unexpected second field: %#v", fields[1])
+	}
+}
+
+func TestImportDocumentWritesSQLiteTables(t *testing.T) {
+	dbPath := filepath.Join(t.TempDir(), "import.sqlite")
+	db, err := sql.Open("sqlite", dbPath)
+	if err != nil {
+		t.Fatalf("open sqlite: %v", err)
+	}
+	defer db.Close()
+
+	xml := `<?xml version="1.0" encoding="UTF-8"?>
+<w:document xmlns:w="http://schemas.openxmlformats.org/wordprocessingml/2006/main">
+  <w:body>
+    <w:tbl>
+      <w:tr><w:tc><w:tcPr><w:gridSpan w:val="2"/></w:tcPr><w:p><w:r><w:t>报备表</w:t></w:r></w:p></w:tc></w:tr>
+      <w:tr><w:tc><w:p><w:r><w:t>序号</w:t></w:r></w:p></w:tc><w:tc><w:p><w:r><w:t>事件过程</w:t></w:r></w:p></w:tc></w:tr>
+      <w:tr><w:tc><w:p><w:r><w:t>1</w:t></w:r></w:p></w:tc><w:tc><w:p><w:r><w:t>客户反映设备位置不合理</w:t></w:r></w:p></w:tc></w:tr>
+    </w:tbl>
+  </w:body>
+</w:document>`
+
+	docxPath := filepath.Join(t.TempDir(), "sample.docx")
+	if err := writeTestDocx(docxPath, []byte(xml)); err != nil {
+		t.Fatalf("write test docx: %v", err)
+	}
+
+	if err := importDocument(context.Background(), db, docxPath); err != nil {
+		t.Fatalf("importDocument: %v", err)
+	}
+
+	var cells int
+	if err := db.QueryRow(`SELECT COUNT(*) FROM cells`).Scan(&cells); err != nil {
+		t.Fatalf("count cells: %v", err)
+	}
+	if cells != 5 {
+		t.Fatalf("got %d cells, want 5", cells)
+	}
+
+	var value string
+	if err := db.QueryRow(`SELECT value FROM fields WHERE key = '事件过程'`).Scan(&value); err != nil {
+		t.Fatalf("query field: %v", err)
+	}
+	if value != "客户反映设备位置不合理" {
+		t.Fatalf("field value = %q", value)
+	}
+}
+
+func writeTestDocx(path string, documentXML []byte) error {
+	f, err := os.Create(path)
+	if err != nil {
+		return err
+	}
+	defer f.Close()
+
+	zw := zip.NewWriter(f)
+	defer zw.Close()
+
+	w, err := zw.Create("word/document.xml")
+	if err != nil {
+		return err
+	}
+	_, err = w.Write(documentXML)
+	return err
+}