处理日期时间值

mssql-python 驱动处理 SQL Server 的日期和时间类型，并将其映射到 Python datetime 的模块对象。 SQL Server 提供了多种日期和时间类型，精度和功能各异：

SQL 类型	Python 类型	范围	精准率
日期	`datetime.date`	0001-01-01 到 9999-12-31	一天
time	`datetime.time`	00:00:00.0000000至23:59:59.9999999	100 纳秒
datetime	`datetime.datetime`	1753-01-01 至 9999-12-31	3.33毫秒
datetime2	`datetime.datetime`	0001-01-01 到 9999-12-31	100 纳秒
smalldatetime	`datetime.datetime`	1900-01-01 至 2079-06-06	1 分钟
datetimeoffset	`datetime.datetime` （与 tzinfo 合作）	0001-01-01 到 9999-12-31	100纳秒+时区

插入日期时间值

使用 Python datetime 对象

以下示例使用Python模块datetime将不同的日期和时间类型插入到SQL Server中：

from datetime import datetime, date, time
import mssql_python

conn = mssql_python.connect(connection_string)
cursor = conn.cursor()

# Create a temp table with date, time, and datetime columns
cursor.execute("""
    CREATE TABLE #Events (
        EventDate DATE,
        StartTime TIME,
        LogTime DATETIME2
    )
""")

# Insert date
cursor.execute(
    "INSERT INTO #Events (EventDate) VALUES (%(date)s)",
    {"date": date(2024, 12, 25)}
)

# Insert time
cursor.execute(
    "INSERT INTO #Events (StartTime) VALUES (%(time)s)",
    {"time": time(14, 30, 0)}
)

# Insert datetime
cursor.execute(
    "INSERT INTO #Events (LogTime) VALUES (%(ts)s)",
    {"ts": datetime(2024, 3, 15, 14, 30, 45)}
)

conn.commit()

插入当前时间戳

您可以使用 datetime.now() 或 SQL Server 的 GETDATE() 函数插入当前日期和时间：

from datetime import datetime

# Create temp table for the demo
cursor.execute("CREATE TABLE #Logs (Message NVARCHAR(200), CreatedAt DATETIME2)")

# Python current time
now = datetime.now()
cursor.execute("INSERT INTO #Logs (Message, CreatedAt) VALUES (%(msg)s, %(ts)s)",
               {"msg": "Event occurred", "ts": now})

# Or use SQL Server's GETDATE()
cursor.execute("INSERT INTO #Logs (Message, CreatedAt) VALUES (%(msg)s, GETDATE())",
               {"msg": "Event occurred"})
conn.commit()

高精度日期时间2

对于更高精度的时间戳，请使用 datetime2 支持100纳秒精度的类型：

from datetime import datetime

# Create temp table with a datetime2(7) column
cursor.execute("CREATE TABLE #PreciseLogs (EventTime DATETIME2(7))")

# Microsecond precision (Python supports up to microseconds)
precise_time = datetime(2024, 3, 15, 14, 30, 45, 123456)

cursor.execute(
    "INSERT INTO #PreciseLogs (EventTime) VALUES (%(ts)s)",  # datetime2(7) column
    {"ts": precise_time}
)
conn.commit()

获取日期时间值

获取日期和时间

从 SQL Server 获取 datetime 值时，它们会自动转换为 Python datetime 对象：

from datetime import date, time, datetime

# Create and populate a temp table with date, time, and datetime columns
cursor.execute("""
    CREATE TABLE #Events (
        ID INT,
        EventDate DATE,
        StartTime TIME,
        CreatedAt DATETIME2
    )
""")
cursor.execute(
    "INSERT INTO #Events (ID, EventDate, StartTime, CreatedAt) VALUES (1, %(d)s, %(t)s, %(dt)s)",
    {"d": date(2024, 12, 25), "t": time(14, 30, 0), "dt": datetime(2024, 3, 15, 14, 30, 45)}
)
conn.commit()

cursor.execute("SELECT EventDate, StartTime, CreatedAt FROM #Events WHERE ID = 1")
row = cursor.fetchone()

print(type(row.EventDate))   # <class 'datetime.date'>
print(type(row.StartTime))   # <class 'datetime.time'>
print(type(row.CreatedAt))   # <class 'datetime.datetime'>

print(row.EventDate)    # 2024-12-25
print(row.StartTime)    # 14:30:00
print(row.CreatedAt)    # 2024-03-15 14:30:45

访问日期时间组件

Python datetime 对象提供属性以访问各个日期和时间组件：

cursor.execute("SELECT OrderDate FROM Sales.SalesOrderHeader WHERE SalesOrderID = 43659")
row = cursor.fetchone()
dt = row.OrderDate

# Date components
print(dt.year)
print(dt.month)
print(dt.day)

# Time components
print(dt.hour)
print(dt.minute)
print(dt.second)
print(dt.microsecond)

时区

包含偏移信息的日期时间与不包含偏移信息的日期时间

Python 的 datetime 对象要么是感知偏移量的（具有 tzinfo），要么是朴素的（没有时区信息）。 SQL Server 有这两种类型的列：

SQL Server类型	期望	商店时区？
datetime，datetime2，smalldatetime	无偏移量	否
datetimeoffset	偏移感知	是的

将带时区信息的日期时间值发送到 datetime2 列是可行的，但时区信息会被静默丢弃。将不含时区信息的日期时间发送到 datetimeoffset 列时，会将偏移量设为 UTC (+00:00)。明确说明你发送的邮件类型：

以下示例演示了如何创建和使用无偏移信息的日期时间和带偏移信息的日期时间：

from datetime import datetime, timezone, timedelta

# Create temp tables: one datetime2 column and one datetimeoffset column
cursor.execute("CREATE TABLE #Logs (LogTime DATETIME2)")
cursor.execute("CREATE TABLE #GlobalEvents (EventTime DATETIMEOFFSET)")

# Offset-naive - use for datetime2 columns
naive_dt = datetime(2024, 3, 15, 14, 30, 45)
cursor.execute(
    "INSERT INTO #Logs (LogTime) VALUES (%(log_time)s)",  # datetime2 column
    {"log_time": naive_dt}
)

# Offset-aware - use for datetimeoffset columns
eastern = timezone(timedelta(hours=-5))
aware_dt = datetime(2024, 3, 15, 14, 30, 45, tzinfo=eastern)
cursor.execute(
    "INSERT INTO #GlobalEvents (EventTime) VALUES (%(event_time)s)",  # datetimeoffset column
    {"event_time": aware_dt}
)
conn.commit()

要在两者之间进行转换：

from datetime import datetime, timezone

# Make naive datetime timezone-aware
naive = datetime(2024, 3, 15, 14, 30)
aware = naive.replace(tzinfo=timezone.utc)

# Strip timezone from aware datetime
aware = datetime.now(timezone.utc)
naive = aware.replace(tzinfo=None)

DateTimeOffset 类型

SQL Server 存储datetimeoffset时区信息。以下示例演示了如何插入和检索带时区信息的日期时间值：

from datetime import datetime, timezone, timedelta

# Create temp table with a datetimeoffset column
cursor.execute("CREATE TABLE #GlobalEvents (ID INT, EventTime DATETIMEOFFSET)")

# Create timezone-aware datetime
eastern = timezone(timedelta(hours=-5))
dt_eastern = datetime(2024, 3, 15, 14, 30, tzinfo=eastern)

cursor.execute(
    "INSERT INTO #GlobalEvents (ID, EventTime) VALUES (1, %(ts)s)",  # datetimeoffset column
    {"ts": dt_eastern}
)
conn.commit()

# Retrieve timezone-aware datetime
cursor.execute("SELECT EventTime FROM #GlobalEvents WHERE ID = 1")
row = cursor.fetchone()
print(row.EventTime)          # 2024-03-15 14:30:00-05:00
print(row.EventTime.tzinfo)   # UTC-05:00

跨时区转换

使用 Python 的时区方法将 datetimeoffset 值在不同时区表示之间转换：

from datetime import datetime, timezone, timedelta

# Create and populate a temp table with a datetimeoffset column
cursor.execute("CREATE TABLE #GlobalEvents (EventTime DATETIMEOFFSET)")
eastern = timezone(timedelta(hours=-5))
cursor.execute(
    "INSERT INTO #GlobalEvents (EventTime) VALUES (%(ts)s)",
    {"ts": datetime(2024, 3, 15, 14, 30, tzinfo=eastern)}
)
conn.commit()

# Retrieve as-stored
cursor.execute("SELECT EventTime FROM #GlobalEvents")
row = cursor.fetchone()
stored_time = row.EventTime  # Has timezone info

# Convert to UTC
utc_time = stored_time.astimezone(timezone.utc)
print(utc_time)

# Convert to local timezone
local_time = stored_time.astimezone()  # System timezone
print(local_time)

SQL中的时区

SQL Server 的 AT TIME ZONE 子句可在查询中在不同时区之间转换 datetimeoffset 值：

from datetime import datetime, timezone, timedelta

# Create and populate a temp table with a datetimeoffset column
cursor.execute("CREATE TABLE #GlobalEvents (EventTime DATETIMEOFFSET)")
eastern = timezone(timedelta(hours=-5))
cursor.execute(
    "INSERT INTO #GlobalEvents (EventTime) VALUES (%(ts)s)",
    {"ts": datetime(2024, 3, 15, 14, 30, tzinfo=eastern)}
)
conn.commit()

# Convert timezone in SQL Server (2016+)
cursor.execute("""
    SELECT 
        EventTime,
        EventTime AT TIME ZONE 'Pacific Standard Time' AS PacificTime,
        EventTime AT TIME ZONE 'UTC' AS UTCTime
    FROM #GlobalEvents
""")

for row in cursor:
    print(f"Original: {row.EventTime}")
    print(f"Pacific: {row.PacificTime}")
    print(f"UTC: {row.UTCTime}")

日期运算

计算日期差异

通过减去日期时间对象计算两个日期之间的天数：

from datetime import timedelta

cursor.execute("SELECT OrderDate, ShipDate FROM Sales.SalesOrderHeader WHERE SalesOrderID = 43659")
row = cursor.fetchone()

# Days between dates
if row.ShipDate and row.OrderDate:
    difference = row.ShipDate - row.OrderDate
    print(f"Shipped in {difference.days} days")

添加区间

使用 timedelta 从日期时间中添加或减去天数、月数或其他时间间隔：

from datetime import datetime, timedelta

# Create and populate a temp table
cursor.execute("""
    CREATE TABLE #Subscriptions (
        ID INT,
        CreatedAt DATETIME2,
        ExpiresAt DATETIME2
    )
""")
cursor.execute(
    "INSERT INTO #Subscriptions (ID, CreatedAt) VALUES (1, %(created)s)",
    {"created": datetime(2024, 1, 1, 12, 0, 0)}
)
conn.commit()

cursor.execute("SELECT CreatedAt FROM #Subscriptions WHERE ID = 1")
row = cursor.fetchone()

# Calculate expiration (30 days from creation)
expiration = row.CreatedAt + timedelta(days=30)
print(f"Expires: {expiration}")

# Update with calculated date
cursor.execute(
    "UPDATE #Subscriptions SET ExpiresAt = %(exp)s WHERE ID = %(id)s",
    {"exp": expiration, "id": 1}
)
conn.commit()

SQL 中的 DATEADD

SQL Server 的DATEADD函数在查询中直接执行日期算术：

# Use SQL Server date functions
cursor.execute("""
    SELECT 
        OrderDate,
        DATEADD(day, 30, OrderDate) AS DueDate,
        DATEADD(month, 1, OrderDate) AS NextMonth
    FROM Sales.SalesOrderHeader
""")

设置值格式

显示格式

使用 strftime() 和格式说明符格式化日期时间值以供显示：

from datetime import datetime

cursor.execute("SELECT TOP(10) OrderDate FROM Sales.SalesOrderHeader")

for row in cursor:
    # strftime formatting
    print(row.OrderDate.strftime("%Y-%m-%d"))           # 2024-03-15
    print(row.OrderDate.strftime("%B %d, %Y"))          # March 15, 2024
    print(row.OrderDate.strftime("%m/%d/%Y %I:%M %p"))  # 03/15/2024 02:30 PM

从字符串中解析

使用以下方法strptime()将字符串表示转换为 Python 日期时间对象：

from datetime import datetime

# Create temp table for the demo
cursor.execute("CREATE TABLE #Events (EventDate DATE)")

# Parse incoming date strings
date_string = "2024-03-15"
parsed_date = datetime.strptime(date_string, "%Y-%m-%d").date()

cursor.execute(
    "INSERT INTO #Events (EventDate) VALUES (%(date)s)",
    {"date": parsed_date}
)
conn.commit()

ISO 8601 格式

ISO 8601格式对于API和JSON序列化非常有用。使用 isoformat() 和 fromisoformat() 进行转换：

from datetime import datetime

# Create and populate a temp table
cursor.execute("CREATE TABLE #Logs (ID INT, CreatedAt DATETIME2)")
cursor.execute(
    "INSERT INTO #Logs (ID, CreatedAt) VALUES (1, %(ts)s)",
    {"ts": datetime(2024, 3, 15, 14, 30, 45)}
)
conn.commit()

cursor.execute("SELECT CreatedAt FROM #Logs WHERE ID = 1")
row = cursor.fetchone()

# ISO format for APIs/JSON
iso_string = row.CreatedAt.isoformat()
print(iso_string)  # 2024-03-15T14:30:45

# Parse ISO format
dt = datetime.fromisoformat("2024-03-15T14:30:45")

常见模式

按日期范围查询

通过参数化起始日期和结束日期，查询特定日期范围内的记录：

from datetime import date, datetime

# Query by date range
start_date = date(2024, 1, 1)
end_date = date(2024, 12, 31)

cursor.execute("""
    SELECT SalesOrderID, OrderDate, TotalDue
    FROM Sales.SalesOrderHeader
    WHERE OrderDate >= %(start)s AND OrderDate < %(end)s
""", {"start": start_date, "end": end_date})

查询今日记录

通过将 datetime 列转换为日期并与当天日期进行比较，查询今天创建的记录：

from datetime import date

# Create and populate a temp table with a row logged today
cursor.execute("CREATE TABLE #Logs (LogTime DATETIME2)")
cursor.execute("INSERT INTO #Logs (LogTime) VALUES (GETDATE())")
conn.commit()

cursor.execute("""
    SELECT * FROM #Logs 
    WHERE CAST(LogTime AS DATE) = %(today)s
""", {"today": date.today()})

# Or using SQL Server function
cursor.execute("""
    SELECT * FROM #Logs 
    WHERE CAST(LogTime AS DATE) = CAST(GETDATE() AS DATE)
""")

处理 NULL 日期

在处理日期时间列之前，请检查是否存在 None 值，因为 NULL 日期在 Python 中会显示为 None：

cursor.execute("SELECT TOP 5 FirstName, ModifiedDate FROM Person.Person")

for row in cursor:
    if row.ModifiedDate is None:
        print(f"{row.FirstName}: Date not on file")
    else:
        age = (date.today() - row.ModifiedDate.date()).days // 365
        print(f"{row.FirstName}: modified {age} years ago")

存储审计时间戳

创建审计栏以跟踪记录的创建和修改时间：

from datetime import datetime

# Create temp table for audit timestamp demo
cursor.execute("""
    CREATE TABLE #AuditOrders (
        ID INT IDENTITY(1,1) PRIMARY KEY,
        CustomerID INT,
        Total DECIMAL(10,2),
        CreatedAt DATETIME2,
        ModifiedAt DATETIME2
    )
""")

def create_order(cursor, customer_id: int, total: float) -> int:
    now = datetime.now()
    cursor.execute("""
        INSERT INTO #AuditOrders (CustomerID, Total, CreatedAt, ModifiedAt)
        VALUES (%(cust)s, %(total)s, %(created)s, %(modified)s);
        SELECT SCOPE_IDENTITY();
    """, {
        "cust": customer_id,
        "total": total,
        "created": now,
        "modified": now
    })
    return cursor.fetchval()

def update_order(cursor, order_id: int, total: float):
    cursor.execute("""
        UPDATE #AuditOrders
        SET Total = %(total)s, ModifiedAt = %(modified)s
        WHERE ID = %(id)s
    """, {
        "total": total,
        "modified": datetime.now(),
        "id": order_id
    })

具体情景

1753年前的日期

用于 datetime2 历史日期（该 datetime 类型仅支持1753年起的日期）。以下示例插入了一个历史日期：

from datetime import datetime

# Create temp table with a datetime2 column
cursor.execute("CREATE TABLE #HistoricalEvents (EventDate DATETIME2, Description NVARCHAR(200))")

# Historical date (works with datetime2, fails with datetime)
historical = datetime(1500, 1, 1)

cursor.execute(
    "INSERT INTO #HistoricalEvents (EventDate, Description) VALUES (%(date)s, %(desc)s)",
    {"date": historical, "desc": "Archived historical record"}
)
conn.commit()

仅含时间的值

使用 Python time 对象存储时间值：

from datetime import time

# Create temp table with time columns
cursor.execute("""
    CREATE TABLE #BusinessHours (
        DayOfWeek NVARCHAR(10),
        OpenTime TIME,
        CloseTime TIME
    )
""")

# Store time of day without date
opening_time = time(9, 0, 0)   # 9:00 AM
closing_time = time(17, 30, 0) # 5:30 PM

cursor.execute("""
    INSERT INTO #BusinessHours (DayOfWeek, OpenTime, CloseTime)
    VALUES (%(day)s, %(open)s, %(close)s)
""", {"day": "Monday", "open": opening_time, "close": closing_time})
conn.commit()

计算工作日

辅助功能可以在跳过周末时计算日期间隔：

from datetime import date, timedelta

def add_business_days(start: date, days: int) -> date:
    """Add business days (skip weekends)."""
    current = start
    remaining = days
    while remaining > 0:
        current += timedelta(days=1)
        if current.weekday() < 5:  # Monday = 0, Friday = 4
            remaining -= 1
    return current

order_date = date(2024, 3, 15)  # Friday
due_date = add_business_days(order_date, 5)  # Skip weekend
print(f"Due date: {due_date}")  # 2024-03-22 (next Friday)

反馈

此页面是否有帮助？

Last updated on 2026-07-22