I am experiencing problem estimating chi-square with tabout command when I svyset the data and use the svy option but that is not the case without the svyset and svy option. I have tried to resolve the problem but to no avail.

Below is the outcome without svyset and svy
Code:
. tabout x symptom if state==1 using table2B.xls, replace nwt(weight) nlab(Number) sebnone h1(nil) clab(freq 95%_CI) f(1 1) stats(chi2)

Table output written to: table2B.xls

x       0       1       Total   Number
        freq    freq    freq    
1       6884.0  922.0   7806.0  7,518
2       4443.0  479.0   4922.0  6,017
Total   11327.0 1401.0  12728.0 13,534

Pearson chi2(1) = 13.3273 Pr = 0.000
Below is the outcome with svyset and svy:
Code:
svyset [iw=weight]

Importance weights: weight
               VCE: linearized
       Single unit: missing
          Strata 1: <one>
   Sampling unit 1: <observations>
             FPC 1: <zero>

. tabout x symptom if state==1 using table2B.xls, replace svy nwt(weight) nlab(Number) sebnone h1(nil) clab(freq 95%_CI) f(1 1) stats(chi2)
Survey results being calculated
----+--- 1 ---+--- 2 ---+--- 3 ---+--- 4 ---+--- 5
.......
Table output written to: table2B.xls

x       0       1       8       Total   Number
        freq    freq    freq    freq    
1       6788.6  729.3   0.0     7517.9  7,791
2       5515.6  500.9   0.0     6016.5  4,920
3       0.0     0.0     0.0     0.0     0
Total   12304.2 1230.2  0.0     13534.4 12,711

Pearson: Uncorrected chi2(4) = .                                        
Design-based F(., .) = . Pr = .
Category 3 of variable x is zero in the cross-tabulation when svyset and svy are used and could be the cause. If it is the cause, how can it be avoided?

How can I get it to estimate chisquare with the svyset and svy option.

Attached is the data. sampledata.dta


Thanks.