这个不同问题的答案是针对修改列的，但我相信，如果你有很多列，并且不想在INSERT语句中手动重新输入大部分列，部分答案也可以产生有用的方法:

https://stackoverflow.com/a/10385666

您可以按照上面链接中描述的方法转储数据库，然后从该转储中获取“create table”语句和“insert”模板，然后按照SQLite FAQ条目“如何从SQLite中现有的表中添加或删除列”中的说明进行操作。(常见问题在本页其他地方有链接。)

基于http://www.sqlite.org/faq.html#q11上的信息在Python中实现。

import sqlite3 as db
import random
import string

QUERY_TEMPLATE_GET_COLUMNS = "PRAGMA table_info(@table_name)"
QUERY_TEMPLATE_DROP_COLUMN = """
  BEGIN TRANSACTION;
  CREATE TEMPORARY TABLE @tmp_table(@columns_to_keep);
  INSERT INTO @tmp_table SELECT @columns_to_keep FROM @table_name;
  DROP TABLE @table_name;
  CREATE TABLE @table_name(@columns_to_keep);
  INSERT INTO @table_name SELECT @columns_to_keep FROM @tmp_table;
  DROP TABLE @tmp_table;
  COMMIT;
"""

def drop_column(db_file, table_name, column_name):
    con = db.connect(db_file)
    QUERY_GET_COLUMNS = QUERY_TEMPLATE_GET_COLUMNS.replace("@table_name", table_name)
    query_res = con.execute(QUERY_GET_COLUMNS).fetchall()
    columns_list_to_keep = [i[1] for i in query_res if i[1] != column_name]
    columns_to_keep = ",".join(columns_list_to_keep)
    tmp_table = "tmp_%s" % "".join(random.sample(string.ascii_lowercase, 10))
    QUERY_DROP_COLUMN = QUERY_TEMPLATE_DROP_COLUMN.replace("@table_name", table_name)\
        .replace("@tmp_table", tmp_table).replace("@columns_to_keep", columns_to_keep)
    con.executescript(QUERY_DROP_COLUMN)
    con.close()

drop_column(DB_FILE, TABLE_NAME, COLUMN_NAME)

这个脚本首先创建随机的临时表，并只插入必要列的数据，除了将要删除的列。然后根据临时表恢复原表，删除临时表。

作为替代:

如果有一个带模式的表

CREATE TABLE person(
  id INTEGER PRIMARY KEY,
  first_name TEXT,
  last_name TEXT,
  age INTEGER,
  height INTEGER
);

你可以使用CREATE TABLE…AS语句，如CREATE TABLE person2 AS SELECT id, first_name, last_name, age FROM person;，即去掉你不想要的列。然后删除原来的person表并重命名新表。

注意，此方法生成的表没有PRIMARY KEY，也没有约束。为了保留这些表，可以使用其他人描述的方法来创建新表，或者使用临时表作为中间表。

SQLite 3.35.0引入了对ALTER TABLE DROP COLUMN的支持。

ALTER TABLE DROP COLUMN语法用于从表中删除现有列。DROP COLUMN命令从表中删除指定的列，并重写整个表以清除与该列相关的数据。DROP COLUMN命令仅在列未被模式的任何其他部分引用、不是PRIMARY KEY且没有UNIQUE约束时才有效。

下面的语法是有效的:

ALTER TABLE <TABLENAME> DROP COLUMN <COLUMNNAME>;
ALTER TABLE <TABLENAME> DROP <COLUMNNAME>;

我重写了@Udinic答案，以便代码自动生成表创建查询。它也不需要ConnectionSource。它还必须在事务中完成此操作。

public static String getOneTableDbSchema(SQLiteDatabase db, String tableName) {
    Cursor c = db.rawQuery(
            "SELECT * FROM `sqlite_master` WHERE `type` = 'table' AND `name` = '" + tableName + "'", null);
    String result = null;
    if (c.moveToFirst()) {
        result = c.getString(c.getColumnIndex("sql"));
    }
    c.close();
    return result;
}

public List<String> getTableColumns(SQLiteDatabase db, String tableName) {
    ArrayList<String> columns = new ArrayList<>();
    String cmd = "pragma table_info(" + tableName + ");";
    Cursor cur = db.rawQuery(cmd, null);

    while (cur.moveToNext()) {
        columns.add(cur.getString(cur.getColumnIndex("name")));
    }
    cur.close();

    return columns;
}

private void dropColumn(SQLiteDatabase db, String tableName, String[] columnsToRemove) {
    db.beginTransaction();
    try {
        List<String> columnNamesWithoutRemovedOnes = getTableColumns(db, tableName);
        // Remove the columns we don't want anymore from the table's list of columns
        columnNamesWithoutRemovedOnes.removeAll(Arrays.asList(columnsToRemove));

        String newColumnNamesSeparated = TextUtils.join(" , ", columnNamesWithoutRemovedOnes);
        String sql = getOneTableDbSchema(db, tableName);
        // Extract the SQL query that contains only columns
        String oldColumnsSql = sql.substring(sql.indexOf("(")+1, sql.lastIndexOf(")"));

        db.execSQL("ALTER TABLE " + tableName + " RENAME TO " + tableName + "_old;");
        db.execSQL("CREATE TABLE `" + tableName + "` (" + getSqlWithoutRemovedColumns(oldColumnsSql, columnsToRemove)+ ");");
        db.execSQL("INSERT INTO " + tableName + "(" + newColumnNamesSeparated + ") SELECT " + newColumnNamesSeparated + " FROM " + tableName + "_old;");
        db.execSQL("DROP TABLE " + tableName + "_old;");
        db.setTransactionSuccessful();
    } catch {
        //Error in between database transaction 
    } finally {
        db.endTransaction();
    }


}

我根据Sqlite推荐的方法写了一个Java实现:

private void dropColumn(SQLiteDatabase db,
        ConnectionSource connectionSource,
        String createTableCmd,
        String tableName,
        String[] colsToRemove) throws java.sql.SQLException {

    List<String> updatedTableColumns = getTableColumns(tableName);
    // Remove the columns we don't want anymore from the table's list of columns
    updatedTableColumns.removeAll(Arrays.asList(colsToRemove));

    String columnsSeperated = TextUtils.join(",", updatedTableColumns);

    db.execSQL("ALTER TABLE " + tableName + " RENAME TO " + tableName + "_old;");

    // Creating the table on its new format (no redundant columns)
    db.execSQL(createTableCmd);

    // Populating the table with the data
    db.execSQL("INSERT INTO " + tableName + "(" + columnsSeperated + ") SELECT "
            + columnsSeperated + " FROM " + tableName + "_old;");
    db.execSQL("DROP TABLE " + tableName + "_old;");
}

为了获得表的列，我使用了“PRAGMA table_info”:

public List<String> getTableColumns(String tableName) {
    ArrayList<String> columns = new ArrayList<String>();
    String cmd = "pragma table_info(" + tableName + ");";
    Cursor cur = getDB().rawQuery(cmd, null);

    while (cur.moveToNext()) {
        columns.add(cur.getString(cur.getColumnIndex("name")));
    }
    cur.close();

    return columns;
}

实际上我在我的博客上写过，你可以在那里看到更多的解释:

http://udinic.wordpress.com/2012/05/09/sqlite-drop-column-support/