Best Practices & Standards for Formatted/Analytical Reporting

    in “Free Characteristics”

    and not in “Rows”

    local views of these queries,

    reducing the number of

    characteristics in the rows will make inserting the queries a bit faster

    YTD, last 90 days, same day

    last year, last 6 months,

    Currency conversion,

    custom/exit variables etc…

    done in BEx only because it

    is usually easier in BEx due to

    offset capability

    Workbooks

    character strings like model

    number for example will be text

    by default but other additive

    numeric attributes need to be

    numeric.

    be formula variable/key

    figures otherwise they are

    text by default. Example:

    Shelf Life in days, Number

    of Scopes Involved, etc…

    values using a matches pattern

    function in Webi. Make the

    numeric values as text.

    This is the opposite of the item

    above.

    a text or character type. Do

    not make it numeric. Example:

    Deal #, Material #, Accounting

    Document Number, etc..

    <<performance may be

    impacted based on this option

    or the one above>>

    pulled in to BO/Crystal

    queries, go to properties

    and select “Use Selection of

    Structure Elements”

    Workbooks

    structure with no data

    choose to purge the last values

    selected before the final save of

    the report

    Workbooks

    not specified in requirements):

    Commas, negative signs,

    decimals?

    on the report with the correct

    formatting of the numbers or

    dates. Make sure currency

    sign, commas, decimals are

    all appropriately applied. Fill

    Empty values with zero, NA,

    or another appropriate value

    instead of leaving blank.

    Workbooks

    or other project logo, color

    scheme, borders, alignment all

    have the same aesthetic look

    and feel for any business area.

    Example : Web reports all look

    and function the same.

    Workbooks

    reports when they are initially

    opened and also to ensure user

    is looking at current data.

    it should be saved with all of the

    data and last selected prompt

    values purged. The report

    should also be saved so that it

    opens on the first report tab

    Workbooks

    this Query” box should be

    checked on the ‘Extended’

    tab of the Query Properties.

    to provide key and text for the

    BO prompt should be multiple

    single value, optional variables.

    Selection Option Variables

    and use Multiple Single Value

    and Range Variables where

    possible. The exception to this

    would be in cases where we

    absolutely need selection option

    variables or we need to use

    authorization based variables.

    right read mode for Queries

    · Read all data (all

    data is read from a

    database and stored in

    user memory space) –

    Used in special cases

    when a majority of the

    users need a given query

    to slice and dice against all

    dimensions

    · Read data during

    navigation (data is read

    from a database only on demand during navigation)

    – Recommended,

    minimizes impact on

    application server

    · Read data during

    navigation and when

    expanding the hierarchy –

    Used for queries involving

    large hierarchies

    navigation” as the default

    and adjust to other options

    if necessary. Use “read all

    data” for non-slice and dice

    queries such as queries

    used for Crystal or Business

    Objects.

    Business Objects

    reporting

    to be processed should be

    minimized with filters. The

    biggest volume reducers,

    should listed first/on top of the

    section

    need a given query to slice and

    dice against all dimensions.

    Reducing the ‘text’ in query will

    also speed up the processing.

    Calculated key figures

    of them during query execution.

    This is very time consuming

    and a high number of then can

    seriously hurt query. Instead,

    define calculated & RKFs as

    much as possible on the Info

    provider level instead of locally

    within the query.

    suppressed

    Key Values” in the query

    properties

    key figures

    Display As Key (or Key and

    Text) and not solely Text for

    characteristics

    they exist in the global filter

    area. Try to remove exclusions

    by subtracting out inclusions.

    as inclusions. Example:

    Company Code >= 2601 *and*

    Company Code <=2599 is

    better than Company Code <>

    2600

    make sure the filter order exists

    with the highest level filter first.

    should come first, this reduces

    records to sort through for the

    remaining filters. Fields with

    the least granularity should be

    restricted first, then the next

    granular field.

    exit variables used in a report.

    global filter whenever possible.

    on multiproviders, restrict to

    InfoProvider in global filter

    whenever possible.

    additional database table joins

    to read data compared to

    those queries against standard

    InfoCubes (InfoProviders), and

    you should therefore hardcode

    the infoprovider in the global

    filter whenever possible to

    eliminate this problem.

    Frontend time whenever

    possible.

    being used, look for other

    options to remove it to

    increase performance.

    Use Constant Selection instead

    of SUMCT and SUMGT within

    formulas.

    BEx Queries/Workbooks

    Do review of

    restrictions in formulas. Do as

    many restrictions as you can

    before calculations.

    restrictions.

    read on Multiproviders.

    if multiple infoproviders are

    involved.

    queries in RSRT.

    exception aggregation on

    calculated key figures.

    slower and should not be used

    unless explicitly needed.

    Business Objects

    possible.

    in production using RSRT

    after changes to statistics,

    consistency changes, or

    aggregates.

    navigational attributes rather

    than hierarchies.

    reading temporary hierarchy

    tables and creates additional

    overhead compared to

    characteristics and navigational

    attributes. Therefore,

    characteristics or navigational

    attributes result in significantly

    better query performance than

    hierarchies, especially as the

    size of the hierarchy (e.g., the

    number of nodes and levels)

    and the complexity of the

    selection criteria increase.

    minimize the number of nodes

    to include in the query results

    query results (even the

    ones that are not needed

    or blank) slows down the

    query processing. The “not

    assigned” nodes in the

    hierarchy should be filtered

    out, and you should use

    a variable to reduce the

    number of hierarchy nodes

    selected.