教程：基于访问控制的ABAP CDS视图权限

Hi!

对每一个CDS视图，我们都可以通过DCL（Data Control Language）定义访问控制。在这篇文章中，我会介绍ABAP CDS视图中非常重要的一面：权限管理。

本文的阐述基于我正在使用的S4/HANA 1610 on NW 7.51.

内容分为五个部分：

标准示例的访问控制。
基于PFCG权限创建一个简单的例子。
带有CUBE数据类别的CDS分析视图。
CDS分析查询视图的访问控制。
权限对象的并集（UNION）或者交集（INTERSECTION）。

本文链接：http://www.cnblogs.com/hhelibeb/p/7427753.html

1. 标准示例的访问控制例子

1) 全访问示例（Full access）

DDL:

@AbapCatalog.sqlViewName: 'DEMO_CDS_FULLACC'

@AccessControl.authorizationCheck: #CHECK

define view demo_cds_auth_fullaccess

  as select from

    scarr

    {

      key carrid,

          carrname,

          currcode,

          url

    };

DCL:

@MappingRole: true

define role demo_cds_role_fullaccess {

  grant select on demo_cds_auth_fullaccess; }

2) 字面条件示例（Literal conditions）

DDL:

@AbapCatalog.sqlViewName: 'DEMO_CDS_LITERAL'

@AccessControl.authorizationCheck: #CHECK

define view demo_cds_auth_literal

 as select from

 scarr

 {

 key carrid,

 carrname,

 currcode,

 url

 };

DCL:

@MappingRole: true

define role demo_cds_role_literal {

  grant select on demo_cds_auth_literal

  where carrid = 'LH'; }

3) PFCG权限示例

DDL:

@AbapCatalog.sqlViewName: 'DEMO_CDS_PFCG'

@AccessControl.authorizationCheck: #CHECK

define view demo_cds_auth_pfcg

 as select from

 scarr

 {

 key carrid,

 carrname,

 currcode,

 url

 };

DCL:

@MappingRole: true

define role demo_cds_role_pfcg {

  grant select on demo_cds_auth_pfcg

  where (carrid) =

  aspect pfcg_auth (s_carrid, carrid, actvt=''); }

权限对象s_carrid可以在事务代码SU21中的BC_C object类下查到。

4) 字面条件和PFCG权限结合示例

DDL:

@AbapCatalog.sqlViewName: 'DEMO_CDS_LITPFCG'

@AccessControl.authorizationCheck: #CHECK

define view demo_cds_auth_lit_pfcg

 as select from

 scarr

 {

 key carrid,

 carrname,

 currcode,

 url

 };

DCL:

@MappingRole: true

define role demo_cds_role_lit_pfcg {

  grant select on demo_cds_auth_lit_pfcg

  where (carrid) =

  aspect pfcg_auth (s_carrid, carrid, actvt='') and

         currcode = 'EUR'; }

5) 继承权限示例

DDL:

@AbapCatalog.sqlViewName: 'DEMO_CDS_INH'

@AccessControl.authorizationCheck: #CHECK

define view demo_cds_auth_inherited

  as select from

    demo_cds_auth_lit_pfcg

    {

      key carrid,

          carrname,

          currcode,

          url

    };

DCL:

@MappingRole: true

define role demo_cds_role_inherited {

  grant select on demo_cds_auth_inherited

               inherit demo_cds_role_lit_pfcg or currcode = 'USD'; }

在这个例子会显示USD和EUR类型货币的记录。

6) 根据当前用户的权限控制示例

DDL:

@AbapCatalog.sqlViewName: 'DEMO_CDS_USR'

@AccessControl.authorizationCheck: #CHECK

define view demo_cds_auth_user

  as select from

    abdocmode

    {

      key uname,

      key langu,

          flag

    };

DCL:

@MappingRole: true

define role demo_cds_role_user {

  grant select on demo_cds_auth_user

    where

      uname ?= aspect user; }

2. 基于PFCG权限创建一个简单的例子

复制以下代码，创建我们自己的CDS视图：

@AbapCatalog.sqlViewName: 'ZDEMO_CDS_PFCG'

@AccessControl.authorizationCheck: #CHECK

@EndUserText.label: 'Demo access pfcg'

define view Zdemo_Access_Pfcg as select from scarr

 {

 key carrid,

 carrname,

 currcode,

 url

 };

3，现在，如果在HANA Studio中打开数据预览，我们将可以看到所有记录。访问控制目前还不存在。

2，在SU21创建我们自己的自定义权限对象：

对于每个对象定义权限字段和活动字段，加入允许活动“03 显示”。在本示例中，我们要在ZS_CONNID中添加字段CARRID和CONNID。

3，为ZS_CARRID创建数据控制。

@MappingRole: true

define role zdemo_access_pfcg {

  grant select on Zdemo_Access_Pfcg

  where (carrid) =

  aspect pfcg_auth (zs_carrid, carrid, actvt=''); }

4，在PFCG中创建一个新的角色，在这里添加刚刚创建的权限对象，定义用户应当看到的基于选择字段的数据。不要忘记生成配置。为我们的用户分配角色。

在第一个示例中，我们只使用ZS_CARRID。在文章的后面，我们会用到其它的对象。

5，回到HANA Studio来测试权限。打开我们的CDS视图的数据预览：

现在我们只看到了定义好的航空公司（CARRID）字段的记录。

注意：

如果在ABAP字典（SE11）中打开视图，结果会是全部数据记录。
如果在DDL中修改注解为如下内容，并激活CDS视图，我们将可以再次在数据预览中看到全部数据。这意味着检查已经关闭。

@AccessControl.authorizationCheck: #NOT_ALLOWED

结论：在一个从数据库表中查询数据的简单例子中，我们看到了访问控制是如何工作的。下面讲讲CDS分析视图。

3. 带有CUBE数据类别的CDS分析视图

1，通过复制已有的内容创建我们自己的CDS视图。这是一个带有CUBE数据分类的CDS视图（译注：代码框出了点问题，大家凑合看下..）：

@AbapCatalog.sqlViewName: 'Z05_CFLIGHTAQ'                       // Name of the CDS database view in the ABAP Repository

@AccessControl.authorizationCheck: #CHECK              // CDS authorizations, controls the authorization check. In S4H410 not required

@EndUserText.label: 'Available Flights'                         // Translatable short text. Max 60characters. Text label is exposed to Analytica tools and the OData service

@VDM.viewType: #CONSUMPTION                                     // This is a CONSUMPTION view

@Analytics.query: true                                          // By tagging the CDS view as an analytical query it will be exposed to the analytic manager

@OData.publish: true                                            // Generates a suitable OData service, that will use the analytical query, when the CDS entity is activated

define view Z05_C_FlightByAirportQuery as select from Z05_I_FlightByAirport     // A analytical query CDS is implemented using a query select from CDS view Z00_I_FlightByAirport

                                                                                // Take care with OData publishing the max. lenght is 26 characters

{

    @AnalyticsDetails.query.axis: #ROWS                         // Defines the default row/colums apperance for the column Airline

    Z05_I_FlightByAirport.Airline,                              // Use the column Airline

    @AnalyticsDetails.query.axis: #ROWS                         // Defines the default row/colums apperance for the column FlightConnection

    Z05_I_FlightByAirport.FlightConnection,                     // Use the column FlightConnection

    @AnalyticsDetails.query.axis: #ROWS                         // Defines the default row/colums apperance for the column FlightDate

    Z05_I_FlightByAirport.FlightDate,                           // Use the column FlightDate

    @Consumption.filter: {selectionType: #SINGLE, multipleSelections: false, mandatory: false }  // Creates a mandatory filter on the values in the field AirportFrom

    @AnalyticsDetails.query.axis: #ROWS                         // Defines the default row/colums apperance for the column AirportFrom

    @EndUserText.label: 'Departure Airport'                     // Add an human readable enduser label to make sure that we can differentiate between AirportFrom and AirportTo

    Z05_I_FlightByAirport.AirportFrom,                          // Use the column AirportFrom

    @Consumption.filter: {selectionType: #SINGLE, multipleSelections: false, mandatory: false } //  Creates an optional filter on the values in the field AirportTo

    @AnalyticsDetails.query.axis: #ROWS                         // Defines the default row/colums apperance for the column AirportTo

    @EndUserText.label: 'Arrival Airport'                       // Add an human readable enduser label to make sure that we can differentiate between AirportFrom and AirportTo

    Z05_I_FlightByAirport.AirportTo,                            // Use the column AirportTo

    Z05_I_FlightByAirport.Currency,                             // Use the column Currency

    Z05_I_FlightByAirport.AircraftType,                         // Use the column AircraftType

    @AnalyticsDetails.query.axis: #COLUMNS                      // Defines the default row/colums apperance for the column FlightPrice

    Z05_I_FlightByAirport.FlightPrice,                          // Use the column FlightPrice

    Z05_I_FlightByAirport.MaximumNumberOfSeats,                 // Use the column MaximumNumberOfSeats

    Z05_I_FlightByAirport.NumberOfOccupiedSeats,                // Use the column NumberOfOccupiedSeats

    @DefaultAggregation: #FORMULA                               // Important to know for formular placement is evaluation time. Inside the final query, the evaluation is done after the flightbyairport

                                                                // view aggragation, so it's not on a very detailed level or even row level, but at the aggragate level. This is important for avarages

                                                                // as they cannot be evaluated at the detail level

    @EndUserText.label: 'Available Seats'

    @AnalyticsDetails.query.axis: #COLUMNS                      // Defines the default row/colums apperance for the column NumberOfAvailableSeats

    Z05_I_FlightByAirport.MaximumNumberOfSeats - Z05_I_FlightByAirport.NumberOfOccupiedSeats as NumberOfAvailableSeats  // this is a formular (calculated column)

}

2，在访问控制中进行定义：

@EndUserText.label: 'Role for Z05_I_FLIGHTBYAIRPORT'

@MappingRole: true

define role Z05_ROLE {

    grant select on Z05_I_FlightByAirport

    where ( Airline ) =

    aspect pfcg_auth (  ZS_CARRID,

                        CARRID,

                        actvt = '' );

}

3，在文章的第2部分，我们在权限对象中添加了ZS_CARRID。在HANA Studio的数据预览中检查结果。行数是530.

4，在事务代码RSRT中检查结果，行数也是530。结果相同。

5，在BO Analysis for Excel中检查结果。结果是相同的，对用户而言，只有选中的航空公司可以被访问。

注意：没有AF航空公司的业务数据，这是上面的屏幕未显示相关数据的原因。

4. CDS分析查询视图的访问控制

1，在第3部分的CUBE CDS中创建一个分析查询视图。

@AbapCatalog.sqlViewName: 'Z05_CFLIGHTAQ'                       // Name of the CDS database view in the ABAP Repository

@AccessControl.authorizationCheck: #CHECK              // CDS authorizations, controls the authorization check. In S4H410 not required

@EndUserText.label: 'Available Flights'                         // Translatable short text. Max 60characters. Text label is exposed to Analytica tools and the OData service

@VDM.viewType: #CONSUMPTION                                     // This is a CONSUMPTION view

@Analytics.query: true                                          // By tagging the CDS view as an analytical query it will be exposed to the analytic manager

@OData.publish: true                                            // Generates a suitable OData service, that will use the analytical query, when the CDS entity is activated

define view Z05_C_FlightByAirportQuery as select from Z05_I_FlightByAirport     // A analytical query CDS is implemented using a query select from CDS view Z00_I_FlightByAirport

                                                                                // Take care with OData publishing the max. lenght is 26 characters

{

    @AnalyticsDetails.query.axis: #ROWS                         // Defines the default row/colums apperance for the column Airline

    Z05_I_FlightByAirport.Airline,                              // Use the column Airline

    @AnalyticsDetails.query.axis: #ROWS                         // Defines the default row/colums apperance for the column FlightConnection

    Z05_I_FlightByAirport.FlightConnection,                     // Use the column FlightConnection

    @AnalyticsDetails.query.axis: #ROWS                         // Defines the default row/colums apperance for the column FlightDate

    Z05_I_FlightByAirport.FlightDate,                           // Use the column FlightDate

    @Consumption.filter: {selectionType: #SINGLE, multipleSelections: false, mandatory: false }  // Creates a mandatory filter on the values in the field AirportFrom

    @AnalyticsDetails.query.axis: #ROWS                         // Defines the default row/colums apperance for the column AirportFrom

    @EndUserText.label: 'Departure Airport'                     // Add an human readable enduser label to make sure that we can differentiate between AirportFrom and AirportTo

    Z05_I_FlightByAirport.AirportFrom,                          // Use the column AirportFrom

    @Consumption.filter: {selectionType: #SINGLE, multipleSelections: false, mandatory: false } //  Creates an optional filter on the values in the field AirportTo

    @AnalyticsDetails.query.axis: #ROWS                         // Defines the default row/colums apperance for the column AirportTo

    @EndUserText.label: 'Arrival Airport'                       // Add an human readable enduser label to make sure that we can differentiate between AirportFrom and AirportTo

    Z05_I_FlightByAirport.AirportTo,                            // Use the column AirportTo

    Z05_I_FlightByAirport.Currency,                             // Use the column Currency

    Z05_I_FlightByAirport.AircraftType,                         // Use the column AircraftType

    @AnalyticsDetails.query.axis: #COLUMNS                      // Defines the default row/colums apperance for the column FlightPrice

    Z05_I_FlightByAirport.FlightPrice,                          // Use the column FlightPrice

    Z05_I_FlightByAirport.MaximumNumberOfSeats,                 // Use the column MaximumNumberOfSeats

    Z05_I_FlightByAirport.NumberOfOccupiedSeats,                // Use the column NumberOfOccupiedSeats

    @DefaultAggregation: #FORMULA                               // Important to know for formular placement is evaluation time. Inside the final query, the evaluation is done after the flightbyairport

                                                                // view aggragation, so it's not on a very detailed level or even row level, but at the aggragate level. This is important for avarages

                                                                // as they cannot be evaluated at the detail level

    @EndUserText.label: 'Available Seats'

    @AnalyticsDetails.query.axis: #COLUMNS                      // Defines the default row/colums apperance for the column NumberOfAvailableSeats

    Z05_I_FlightByAirport.MaximumNumberOfSeats - Z05_I_FlightByAirport.NumberOfOccupiedSeats as NumberOfAvailableSeats  // this is a formular (calculated column)

}

2，在HANA Studio中进行数据预览，行数还是4894。看起来CDS分析查询没有使用到Cube CDS视图权限，但是事实并非如此。你并不需要为分析查询CDS视图创建额外的访问控制。

3，在Excel中检查RSRT或者BO分析的结果。结果表明Cube CDS视图的权限在分析查询中起到了作用。

注意：在分析查询定义中不需要创建任何变量，就像我们在带有权限的BEx查询中那样。

4，修改Cube CDS视图，添加权限对象ZS_CONNID而非ZS_CARRID。

@EndUserText.label: 'Role for Z05_I_FLIGHTBYAIRPORT'

@MappingRole: true

define role Z05_ROLE {

    grant select on Z05_I_FlightByAirport

     where ( FlightConnection) = aspect pfcg_auth (  ZS_CONNID,

                                                     CONNID,

                                                     actvt = '' );

}

分析查询结果变得严格了（在第2部分的第4步可以看到ZS_CONNID的定义）.

现在结果的行数是212.

5. 权限的并集（UNION）和交集（INTERSECTION）

1，通过“AND”取权限的交集。这里定义了一个新的权限“ZS_FLDAT”，它只包含3天的范围（2015.02.04 - 2015.02.06）。修改DCL，增加交集：

@EndUserText.label: 'Role for Z05_I_FLIGHTBYAIRPORT'

@MappingRole: true

define role Z05_ROLE {

    grant select on Z05_I_FlightByAirport

     where ( Airline) =

            aspect pfcg_auth (  ZS_CARRID,

                                CARRID,

                                actvt = '' ) AND

           (FlightDate ) =

            aspect pfcg_auth (  ZS_FLDAT,

                                FLTDATE,

                                actvt = '' );

}

2，通过“OR”取并集：

@EndUserText.label: 'Role for Z05_I_FLIGHTBYAIRPORT'

@MappingRole: true

define role Z05_ROLE {

    grant select on Z05_I_FlightByAirport

     where ( Airline) =

            aspect pfcg_auth (  ZS_CARRID,

                                CARRID,

                                actvt = '' ) OR

           ( FlightDate ) =

            aspect pfcg_auth (  ZS_FLDAT,

                                FLTDATE,

                                actvt = '' );

}

3，如果在一个权限对象中添加这两个字段，那结果就类似于交集：

@EndUserText.label: 'Role for Z05_I_FLIGHTBYAIRPORT'

@MappingRole: true

define role Z05_ROLE {

    grant select on Z05_I_FlightByAirport

     where ( Airline, FlightDate) =

            aspect pfcg_auth (  ZS_NEW,

                                CARRID,

                                FLTDATE,

                                actvt = '' );

注意：不要忘记在Cube CDS视图的层级定义权限，而非分析视图层级。如果你在分析查询层级定义了和第5部分相同的权限，那么：

在SAP HANA Studio的数据预览中，结果看起来是对的。
在RSRT, BO Analysis for Excel和其它使用了OLAP引擎的工具中，使用的是Cube CDS视图的权限（如有定义）。

注意：在HANA Studio的数据预览中，分析查询的结果会全部展示。为了纠正这点，可以给分析查询创建以下访问控制：

@MappingRole: true

define role Z05_ROLE_2 {

  grant select on Z05_C_FlightByAirportQuery

               inherit Z05_ROLE; }

结论：你可以为CDS分析视图定义权限的交集或者并集。

本文结束，感谢关注！

英文原文：ABAP CDS views with Authorization based on Access Control