|
|
,
S
TA
R
T
O
F
Q
UA
R
T
E
R
T
he S
TAR
TOFQU
AR
TER
func
tion re
t
urns the da
te of
t
he
st
art
of
the qua
r
t
e
r
.
L
AS
T
D
A
T
E
T
he L
AS
TDA
TE
func
tio
n r
e
t
urns the l
ast
d
ate of
t
he
quarter.
'
Da
t
e
'
S
pe
c
if
ie
s t
he Da
t
e
tabl
e
.
[
Da
t
e
]
S
pe
c
if
ie
s t
he Da
t
e
c
ol
umn in the Da
t
e
tabl
e
.
,
E
ND
O
F
Q
U
A
RT
E
R
T
he EN
DO
FQU
AR
TER
func
tion
'
Da
t
e
'
S
pe
c
if
ie
s t
he Da
t
e
tabl
e
.
[
Da
t
e
]
S
pe
c
if
ie
s t
he Da
t
e
c
ol
umn in the Da
t
e
tabl
e
.
FO
R
M
U
L
A ELEM
ENT
D
ES
C
R
I
PT
IO
N
Usin
g
f
o
r
m
u
la
A
u
t
o
Co
m
p
l
e
t
e
Usin
g
f
o
r
m
u
la
A
u
t
o
Co
m
p
l
e
t
e
Usin
g
m
u
ltiple
fun
ctio
n
s
in
a
f
o
r
m
u
la
Usin
g
m
u
ltiple
fun
ctio
n
s
in
a
f
o
r
m
u
la
F
un
cti
o
ns
B
oth
the form
ul
a b
ar
in the
m
ode
l
d
es
igner
an
d
the fo
r
m
u
la R
ow F
i
lters
w
indow
in t
he R
ole M
anager dialog
box
prov
ide an Au
toC
om
plet
e feat
ur
e. A
utoC
om
plet
e helps you ente
r
a val
id
f
orm
ul
a sy
ntax
by pr
ovi
ding
y
ou w
ith
opt
ions
for each
elem
ent
in
t
he form
u
la.
Y
ou can use form
ul
a A
utoC
om
plete
in the
m
iddle o
f an existing form
ul
a wi
th
n
este
d functions. Th
e
tex
t
im
m
ediate
l
y befo
r
e t
h
e
i
nser
tion po
in
t
is
us
ed
to
display
v
alues in
t
he d
r
op
-
down
l
is
t,
and
al
l of
the t
ext
after the
ins
er
tion point
r
em
ains
unchanged
.
A
utoC
omplete
doe
s
not
add
the
closi
ng p
ar
enthesis
of
f
unctions or automatically
match p
ar
entheses
.
Y
ou
m
us
t
m
ake sur
e t
h
at
each funct
i
on
i
s s
yn
tac
tically
co
r
r
ect
or
y
ou cannot
sav
e or us
e
the fo
r
m
u
la.
Y
ou can nest
fu
nct
i
ons,
m
eani
ng
that you use t
he r
esults
from
one
fun
ct
ion as
an
argum
ent
of anothe
r
fun
ct
ion.
Y
ou can nest
up t
o 6
4
levels
o
f fun
ct
ions
in calculate
d co
l
um
ns
.
H
ow
ev
er
,
nesting can
m
ak
e
i
t
diffi
cult
to
cr
eat
e or
troublesh
oot
form
ul
as
.
M
an
y fun
ct
ions
ar
e
designed t
o b
e used
s
olely
as nested
functions. T
hese funct
i
ons r
et
u
r
n a
tab
l
e,
w
hich cannot
be
dir
ect
l
y sav
ed as a
r
es
ult;
it mus
t
be p
r
ov
ided
as
input
to
a t
able function.
F
or
ex
am
ple,
the
functions S
U
M
X, A
V
E
R
AG
E
X
,
and M
I
N
X
all r
equir
e
a tab
l
e
as
t
h
e
fi
r
st argum
ent.
DA
X
inclu
de
s
functions you can use t
o p
er
for
m
c
al
culations us
ing d
ates and t
im
es
,
create
condit
i
onal v
alues, wor
k
w
ith str
ings, p
er
for
m
lookups based
on relations
hips
,
and t
he ability
to it
er
ate
ov
er a t
able t
o pe
r
for
m
recur
s
iv
e
calculations. I
f you are fam
il
iar
w
ith E
xcel
f
orm
ul
as, many
of t
hese functions w
il
l app
ear
ver
y
sim
i
lar
;
how
ev
er
,
DA
X
for
m
ulas
are differ
ent
in the
foll
ow
ing im
po
r
tant
w
ays
:
A
DA
X
fun
ct
ion alw
ay
s
refer
ences a
complete
colum
n or a t
able. If y
ou
w
ant
to
us
e o
n
ly
p
ar
ticular
v
alues
fr
om
a
table
or
c
olu
m
n,
y
ou c
an add
fi
lters
t
o t
he form
u
la.
I
f you need
to
customi
ze
calculations
o
n
a
r
ow
-
by
-
r
ow
b
as
is
, DA
X
p
r
ov
ides functions that
let
y
ou use t
h
e
cur
r
ent
r
ow
val
ue
or
a related
v
alue as a kind o
f
p
ar
am
ete
r
, t
o p
er
for
m
c
al
culations that
v
ar
y
b
y
c
ontext
. T
o
under
stand
h
ow
t
hes
e
fu
nct
i
ons w
or
k
,
s
ee
C
ontex
t
in t
hi
s ar
ticle.
DA
X
inclu
de
s
man
y fun
ct
ions
t
h
at
r
eturn
a
table,
rather
t
han
a
v
alu
e.
T
he t
able is not
dis
play
ed
in a
D
ate
and time
f
unc
tio
ns
D
ate
and time
f
unc
tio
ns
F
ilt
e
r f
unct
io
ns
F
ilt
e
r f
unct
io
ns
I
nf
o
rmat
io
n
f
unc
t
io
ns
I
nf
o
rmat
io
n
f
unc
t
io
ns
L
o
g
ic
al
f
unc
tio
ns
L
o
g
ic
al
f
unc
t
io
ns
M
athe
mat
ic
al
and
trig
o
no
me
t
ric
f
unct
io
ns
M
athe
mat
ic
al
and
trig
o
no
me
t
ric
f
unct
io
ns
O
t
he
r f
unc
tio
ns
O
t
he
r f
unc
tio
ns
St
atist
ic
al
f
unc
t
io
ns
St
atist
ic
al
f
unc
t
io
ns
T
e
xt f
unc
tio
ns
T
e
xt f
unc
tio
ns
T
ime
-
inte
llig
e
nc
e
f
unct
io
ns
T
ime
-
inte
llig
e
nc
e
f
unct
io
ns
T
ab
le
-
valued
f
unc
tio
ns
T
ab
le
-
valued
f
unc
tio
ns
r
epo
r
ting c
l
ient,
but
is
used t
o p
r
ov
ide
i
nput
to
other functions. F
or example,
y
ou c
an r
et
r
i
eve a
table and
then
count
the d
i
stinct
v
al
ues in it,
or calculate
dy
nam
ic sum
s
ac
r
os
s fil
te
r
ed
tables or colum
ns
.
DA
X
functions in
clude
a var
i
et
y
of
tim
e
-
inte
llige
nc
e
functions. T
hese functions let
y
ou d
efin
e
or
select
dat
e
r
anges,
and pe
r
for
m
dynam
i
c
calculations
b
as
ed
on t
h
ese d
ates or r
ange.
F
or example,
y
ou c
an
c
om
pare
su
m
s acros
s par
allel periods.
T
he d
ate
and t
i
m
e
fu
nctions in DA
X ar
e sim
il
ar
t
o d
ate and
tim
e
fu
nct
i
ons in M
icros
oft E
xc
el. H
owev
er
,
DA
X
functions ar
e b
as
ed
on t
h
e
d
a
te
tim
e
dat
a types used b
y
M
i
cros
oft
S
Q
L
S
er
v
er
.
F
or
mor
e infor
m
ation,
see
D
ate
and t
im
e functions
.
T
he filter functions in DA
X
retur
n spec
i
fic d
ata t
y
pes,
l
oo
k
up
v
alu
es in related
tales, and
fil
te
r
by r
elate
d val
ues. Th
e
lookup funct
i
ons w
or
k by usi
ng t
ables and relations
hips
,
lik
e
a datab
as
e.
T
he filteri
ng funct
i
ons let
y
ou man
ipulate
dat
a cont
ext
to
cr
eat
e d
y
nam
i
c
calculations
. F
or
mor
e
i
nfor
m
ation,
see
F
ilter fun
ct
ions
.
A
n infor
m
ation funct
i
on loo
k
s at
the c
ell
o
r
r
ow
t
hat is prov
ided
as
an
ar
gum
ent and
te
l
ls
you w
het
h
er the
v
alu
e
m
atc
hes
t
he ex
pect
ed
type. F
or
e
xam
ple,
the I
S
E
R
R
O
R
function return
s T
R
U
E
if the
v
alu
e
that
y
ou refer
ence
cont
ai
ns an err
or
. F
or
mor
e infor
m
ation,
see
I
nfor
m
ation
fu
nct
i
ons
.
L
ogical funct
i
ons act
upon an
expres
si
on t
o return
infor
m
ation
about
the valu
es in t
h
e
expres
si
on.
F
or
e
xam
ple,
the
T
R
U
E
funct
i
on let
s
you know
w
het
h
er an ex
pr
ess
ion that
y
ou
ar
e eval
uating
r
etur
ns
a T
R
U
E
v
alue. F
or
mor
e
inf
orm
ation,
s
ee
L
og
i
cal funct
i
ons
.
T
he mathematical functions in DA
X
are ver
y
sim
i
lar
t
o t
h
e E
xce
l m
athematical and
tr
igonometri
c
fu
nct
i
ons. S
om
e
m
inor
differences exist in t
he num
er
ic
data
ty
pe
s
used b
y
DA
X fun
ct
ions
. F
or
mor
e infor
m
ation,
see
M
ath
an
d
tr
ig
functions
.
T
hese functions pe
r
for
m
unique act
ions
t
ha
t
cannot
be
defined b
y
any of t
he cat
egories
most ot
her
functions
belong
to
. F
or m
ore infor
m
ation,
s
ee
O
ther functions
.
DA
X
prov
ides statistical functions that
perfor
m
agg
r
egat
i
ons. I
n ad
dition t
o creating
s
um
s
and
av
er
ages,
or
finding
the
m
in
im
um
and
m
axim
um
v
alues
,
in
DA
X
you can also filter a c
olu
m
n b
efor
e ag
gr
egat
in
g
or
c
r
eate
aggregations
based
on r
elate
d t
ables
.
F
or
mor
e infor
m
ation,
see
Statistical functions
.
T
he t
ext
functions in DA
X
are ver
y
sim
i
lar
t
o t
h
eir
c
ounterpar
ts in E
xc
el. Y
ou can retur
n p
ar
t o
f
a
s
tri
ng,
sear
ch for
te
xt
w
ithin a stri
ng,
or conc
atenate
str
in
g
v
alu
es. DA
X
also prov
ides functions for cont
r
oll
ing t
h
e
for
m
ats for date
s
,
times
,
and num
bers
. F
or
mor
e
i
nfor
m
ation,
see
T
ext
fun
ct
ions
.
T
he t
im
e
-
intelli
ge
nce funct
i
ons prov
ided
i
n DA
X
let
y
ou c
r
eate
calculations
t
hat use b
u
ilt
-
in know
l
ed
ge ab
out
calendars
and
date
s
. B
y us
ing t
i
m
e
an
d
date
r
anges in co
m
bination w
ith agg
r
egat
i
ons or calculations
,
you can b
u
ild
m
eaningful
c
om
pari
sons
acros
s co
m
par
able t
im
e pe
r
i
od
s
for sa
les,
i
nv
ento
r
y
,
an
d
s
o
on. F
or m
ore infor
m
ation,
see
T
im
e
-
intelli
ge
nce funct
i
ons
(
DA
X
)
.
T
her
e
ar
e DA
X functions that
output
tab
l
es,
take t
ables as in
put
, o
r
do
bot
h. B
ecause a
table
can hav
e a
s
ingle
colum
n,
table
-
v
alued funct
i
ons als
o
take sin
gle
colum
ns
as inputs. U
nder
standing how to
us
e t
able
-
v
alued
V
a
ri
a
b
les
VAR
TotalQty = SUM ( Sales[Quantity] )
Return
IF (
TotalQuantity > 1000,
TotalQuantity * 0.95,
TotalQuantity * 1.25
)
Da
ta
typ
es
D
AT
A T
YPE I
N M
O
D
EL
D
AT
A T
YPE I
N D
AX
D
ES
C
R
I
PT
IO
N
W
hol
e N
umbe
r
A 64
bi
t
(
ei
ght
-
bytes
)
inte
ge
r
val
ue
N
umb
ers that have no d
ec
im
al p
la
ce
s.
Integ
ers c
an be
posi
tive o
r
ne
gati
v
e
numb
ers, but must be
whol
e num
be
r
s
be
twee
n
-
9,2
2
3,3
72
,0
36
,85
4
,77
5,8
0
8
(
-
2^
6
3
)
and
9,2
23
,3
72
,03
6
,85
4,7
7
5,8
07
(
2^
6
3
-
1
)
.
functions is
im
portant
for
full
y utili
zing DA
X
for
m
ulas
. DA
X
includes t
he follow
in
g
types of t
able
-
v
alued funct
i
ons:
F
ilter fun
ct
ions
r
et
ur
n
a
colum
n,
table,
or v
alues
related
to
the
cur
r
ent
r
ow
.
A
ggregat
i
on funct
i
ons agg
r
egate
any
e
xpr
ess
ion over
t
he row
s
of
a tab
l
e.
T
im
e
-
intelli
ge
nce funct
i
ons r
et
u
r
n a
table
of dat
es
,
or us
e
a table
of d
ates to
calculate
an aggregat
i
on.
E
xam
ples of
table funct
i
ons include:
F
I
LT
E
R
, A
L
L
, VA
L
U
E
S
, D
I
S
T
I
N
C
T
, R
E
L
A
T
E
DT
A
B
L
E
.
Y
ou can c
r
eate
v
ar
iables
wi
thin an e
xpr
ess
ion by usi
ng
the
VA
R
. VA
R
is te
chn
ically
not
a function,
it's
a
k
eyw
or
d
y
ou
u
se t
o store t
h
e
r
esu
lt
of an ex
pr
ess
ion as a nam
ed
v
ar
iable. That var
ia
ble
can t
h
en
be p
as
s
ed
as
an
ar
gum
ent
to
ot
h
er m
easur
e ex
pr
ess
ions
. F
or
e
xam
ple:
I
n t
h
is
e
xam
ple,
T
ot
al
Q
ty can
then be
pass
ed
as
a nam
ed
v
ar
i
able
to
other exp
r
es
si
ons. V
ar
iables
c
an
be o
f
any
scalar
data
type,
includin
g
tables. U
s
ing var
i
ables in your
DA
X
form
ul
as can b
e incr
ed
i
bly power
fu
l.
Y
ou can impor
t
data
i
nto
a model from
man
y differ
ent
data
s
our
ces t
hat mi
ght
su
pp
or
t d
if
ferent d
ata t
y
pes. W
hen
y
ou
i
m
po
r
t d
ata into
a model,
the
data
i
s conver
ted
t
o one
of t
h
e
tabular
model d
ata t
y
pes. W
hen the
m
ode
l
d
ata is
us
ed
in a calculation,
the
dat
a is
t
h
en
conv
erted
to
a DA
X
d
ata t
y
pe
for
t
h
e
dur
ation and
output
of t
he c
al
culation.
W
hen you create
a DA
X
for
m
ula,
the t
er
m
s us
ed
in
t
he form
ul
a w
ill
aut
om
atically
d
eterm
in
e
the valu
e
data
type
r
et
u
r
ned.
DA
X
su
pp
or
ts the
foll
owi
ng d
ata t
y
pes:
1
,
2
De
ci
mal
N
umbe
r
A 64
bi
t
(
ei
ght
-
bytes
)
rea
l numb
er
R
ea
l numb
ers are
numbe
rs t
hat c
an
have
de
c
ima
l p
lac
e
s.
R
ea
l numb
ers
c
over a wi
de
range
of
v
al
ues:
N
e
gati
v
e val
ues f
rom
-
1.7
9
E +3
08
through
-
2.2
3
E
-
30
8
Ze
ro
Po
sitive
v
al
ues f
rom 2.2
3
E
-
30
8
through
1.7
9
E + 3
08
Ho
we
v
er, the
numb
er of
sig
nif
ic
ant
di
gi
ts is li
mi
ted
to 17
de
ci
mal
di
gi
ts.
Bo
ol
ea
n
Bo
ol
ea
n
Ei
ther a T
rue or Fal
se val
ue.
T
ex
t
S
tring
A Uni
co
de
c
harac
ter da
t
a string
. Can be
string
s, numbe
rs or date
s r
e
prese
nted
i
n
a tex
t
fo
rmat.
Date
Date
/ti
me
Date
s and tim
es i
n an ac
ce
pted
date
-
tim
e rep
rese
nt
ati
on.
V
ali
d d
ates are
all
da
t
e
s
af
ter March 1,
19
0
0.
Currenc
y
Currenc
y
Currenc
y data type
all
ow
s v
al
ues
be
twee
n
-
92
2
,33
7
,20
3,6
85
,4
77
.58
0
8
to
92
2
,33
7
,20
3,6
85
,4
77
.58
0
7
wi
th four
de
c
ima
l di
gi
ts of
f
ix
ed
prec
i
sio
n.
N
/A
Bl
ank
A b
la
nk
i
s a data type
in DAX
that
rep
rese
nts
and
repl
ac
es S
Q
L null
s. Y
ou
c
an cre
ate a
bl
ank b
y
usi
ng the BL
AN
K
f
uncti
on, and
test f
or bl
anks by usin
g
the l
ogi
c
al
f
uncti
on, I
S
BL
AN
K.
D
AT
A T
YPE I
N M
O
D
EL
D
AT
A T
YPE I
N D
AX
D
ES
C
R
I
PT
IO
N
Co
ntex
t
1
,
2
T
abula
r
data mode
l
s als
o include
the
T
a
b
le
dat
a type
as
t
he input o
r
outp
ut t
o man
y DA
X
funct
i
ons. F
or ex
am
ple,
the F
I
LT
E
R
function
takes a t
able as input and
outp
u
ts anot
her
t
able that
co
n
tains only
t
he r
ow
s that
m
eet
t
h
e
fi
lter
cond
i
tions. B
y combini
ng t
able funct
i
ons w
ith agg
r
egat
i
on funct
i
ons,
y
ou can
per
for
m complex c
alcul
ations over
dynam
i
cally
d
efi
ned
data
s
et
s
.
W
hil
e
data
types ar
e t
y
pically
aut
om
atically
set,
it
is
im
portant
to
u
nders
tand
data
types and how they app
ly
,
i
n
-
particular
, t
o DA
X
for
m
ulas
. E
r
r
ors
in form
u
las
o
r
unexpe
cted
r
esults
,
for
e
xam
ple,
ar
e
often c
aus
ed
by
usin
g
a
particular
ope
r
ator
t
hat c
ann
ot
be
us
ed
w
i
th
a dat
a type
s
pec
i
fied
i
n an argum
ent.
F
or
e
xam
ple,
the
for
m
ula,
= 1 & 2
,
r
eturn
s a stri
ng resu
lt
of 1
2
.
T
h
e
for
m
ul
a,
= "1" + "2"
,
how
ev
er
,
r
eturn
s an integ
er
res
ult o
f
3.
C
o
nte
xt
is an im
portant
conce
pt
to
u
nders
tand
w
h
en
cr
eating DA
X
form
ul
as. C
ont
ext
is
w
hat
enables
you t
o
perfor
m
d
y
nam
i
c
an
aly
si
s, as t
he r
esults
o
f a form
ul
a c
h
ange
to
r
eflect
the
cur
r
ent
r
ow
or cell select
i
on
an
d
al
so
any
related
data.
U
nde
r
s
tanding c
ontex
t and
us
in
g
conte
xt
effect
i
v
ely ar
e c
r
itical for
b
ui
lding high
-
perfor
m
ing,
dynam
i
c
ana
ly
ses
,
and for t
r
oubles
hoot
i
ng
pr
oblems
in form
u
las
.
F
orm
ul
as
in t
abular
m
ode
ls
can
be e
v
al
uate
d in a d
i
ffer
ent
cont
ext,
d
epending
on ot
her
de
s
ign elements:
Ro
w c
o
nte
x
t
Ro
w c
o
nte
x
t
= [Freight] + RELATED('Region'[TaxRate])
Mu
ltiple
r
ow
co
n
t
ex
t
Mu
ltiple
r
ow
co
n
t
ex
t
=MAXX(FILTER(Sales,[ProdKey]=EARLIER([Prod
Key])),Sales[OrderQty])
Qu
e
r
y
co
n
t
ex
t
Qu
e
r
y
co
n
t
ex
t
F
ilters
app
l
ied
i
n a P
iv
otT
able or r
epo
r
t
F
ilters
de
fi
ned
w
i
thin a form
ul
a
Relations
hips
spec
i
fied
by
usin
g
s
pec
i
al functions w
ithin a form
u
la
T
her
e
ar
e d
i
ffer
ent
types of c
ontex
t:
r
o
w
co
nte
xt
,
q
ue
r
y co
n
text
,
and
filter
co
nte
xt
.
Row
co
n
text
c
an
b
e t
h
ought
of as "
the c
u
r
r
ent
r
ow
".
I
f y
ou
cr
eate
a form
ul
a in a
calculated
colum
n,
the row
co
n
te
xt
for
t
hat form
ul
a includes t
he v
alues fr
om
all co
l
um
ns
in t
he curr
ent row
. If
t
he t
able is r
elated
to
anothe
r
tab
l
e,
the
cont
ent
al
so includes all
t
he val
ues fr
om
t
he o
ther
t
able t
hat are r
elate
d t
o t
he curr
ent row.
F
or example,
s
uppo
s
e you c
r
eate
a calculate
d c
olu
m
n,
=[Freight] + [Tax]
,
that
adds to
get
h
er v
alues
from
t
w
o
colum
ns, F
r
eight
and T
ax,
fr
om
t
he sam
e t
able. Thi
s for
m
ula automatically
ge
ts only
t
h
e
v
alu
es fr
om the
cur
r
ent
r
ow
in t
he specified
colum
ns
.
Row
cont
ext
als
o fo
l
low
s
any r
elationshi
ps t
h
at
hav
e be
en de
fi
ned
bet
w
ee
n
t
ables, including relations
hips defined
w
ithin
a
calculated
colum
n by usi
ng DA
X
form
u
las
,
to
det
er
m
ine w
hich row
s
in r
elate
d t
ables
are ass
oc
i
ate
d w
ith
the
cur
r
ent
r
ow.
F
or example,
the
foll
ow
ing form
ul
a uses
t
he R
E
L
A
T
E
D
funct
i
on
to fe
tch a
tax
v
alu
e
fr
om
a related
tab
l
e,
based
on
the
r
egion t
hat t
he order w
as
shi
pp
ed t
o.
T
he tax
v
alue is det
er
m
ined
by
usi
ng
the valu
e
for
r
eg
i
on in t
he curr
ent
tab
l
e,
looking up
the
r
egion in t
h
e
r
elated
table,
and
then g
ett
i
ng
the t
ax rate
f
or that
r
egion
fr
om
t
h
e
r
elated
table.
T
his
form
ul
a g
ets the
tax
r
ate
for
t
h
e
cur
r
ent
r
egion fr
om the
R
egion t
able and
adds it
to
the
v
al
ue o
f the F
r
eight
colum
n. In DA
X
form
ul
as, you d
o not
need
to
kn
ow or
spec
if
y the
s
pec
i
fic relations
hip t
hat co
nn
ec
ts the
tables.
DA
X
inclu
de
s
functions t
h
at
iter
ate
calculations over
a
table.
T
hes
e
fu
nctions can
h
ave m
ultiple c
ur
r
ent row
s
,
each
w
ith its own
row
co
ntext
. In
e
s
sence,
these
f
unctions let
y
ou c
r
eate
for
m
ulas
t
h
at
perf
orm
op
er
ations r
ecurs
iv
ely
over
an inner
and
outer loop
.
F
or example,
s
uppo
s
e your
model
contains
a
P
rod
u
c
ts
t
able and
a
S
a
l
e
s
t
able. U
s
ers
m
ight
w
an
t
to
go
thr
ough
the
entir
e
s
ales
t
able,
w
hi
ch is full
o
f
t
r
ans
act
i
ons inv
olv
i
ng
m
u
ltiple
pr
oduct
s
,
and find t
he lar
gest
quantity order
ed
for
e
ach product
in any
one
tran
saction.
W
ith DA
X
you can b
u
ild a
s
in
gle
for
m
ul
a t
hat return
s the
corr
ect
v
alu
e,
and
the resu
lts ar
e
au
t
om
atically
upd
ated
any
t
i
me a user
ad
ds
data t
o t
he t
ables
.
F
or a d
etailed
exam
ple
of this for
m
ula, see
E
A
R
L
I
E
R
.
T
o sum
m
ar
i
ze,
t
he E
A
R
L
I
E
R
fun
ct
ion s
t
or
es the
r
ow
co
n
t
ext
fr
om
t
h
e
operation t
hat prece
ded
the c
ur
r
ent
ope
r
ation. At all t
i
m
es,
the funct
i
on
s
tores in
mem
ory
t
w
o
s
et
s
of
conte
xt:
one
s
et
of c
ontex
t repr
esents t
h
e
cur
r
ent
r
ow
for t
h
e
i
nner
loo
p of
the form
ul
a,
and anot
her
set
of co
ntext
r
ep
r
es
ent
s
the
cur
r
ent
r
ow
for the
outer loop
of
the
for
m
ula. DA
X
automatically
feed
s
valu
es be
tw
een
the t
w
o loop
s
so t
hat you can
cr
eate
co
m
plex ag
gr
egat
es.
Q
u
e
ry
co
nte
xt
refer
s to
the
s
ubset
of dat
a that
is
im
plicitly r
et
r
i
eved fo
r
a form
ul
a. W
hen a user
p
l
aces a
m
easu
r
e
or othe
r
v
alue field into
a P
iv
otT
able o
r
into
a r
epo
r
t b
as
ed
on a t
abular
m
ode
l, t
he engine
exam
ines the
r
ow
and
colum
n head
er
s
,
S
li
ce
r
s
,
and r
ep
or
t
f
ilters
t
o de
term
i
ne t
he c
ontext
. Then, t
he nec
es
sar
y
que
r
i
es ar
e
r
un
ag
ai
ns
t
the
dat
a sour
ce t
o g
et
the c
or
r
ect
subs
et
of d
ata,
m
ake t
he calculations de
f
ined
by
t
he form
u
la,
and t
h
en p
opulate
each
cell in t
he P
iv
otT
able o
r
r
epo
r
t.
T
h
e
s
et
of d
ata t
hat is r
et
r
i
eved is t
h
e
quer
y
contex
t fo
r
each
cell.
剩余509页未读， 继续阅读
-
我的内容管理
快来上传第一个资源
-
我的收益
登录查看自己的收益 -
我的积分
登录查看自己的积分
登录后查看C币余额
-
我的收藏
-
我的下载
-
下载帮助
