A decomposition of relation R into R1,R2,…,Rk is lossless if we can successively join pairs of subrelations such that the common attributes form a superkey of at least one of the subrelations.
### Analysis of Decomposition D1:Constituent subrelations: $R_1(P, Q, S, T)$, $R_2(P, T, X)$, $R_3(Q, Y)$, $R_4(Y, Z, W)$.
Join $R_3(Q, Y)$ and $R_4(Y, Z, W)$:
Common attribute: R3∩R4={Y}.
Given Y→ZW, Y is a candidate key for R4.
Thus, R3⋈R4=R34(Q,Y,Z,W) is a lossless join.
Join $R_1(P, Q, S, T)$ and $R_2(P, T, X)$:
Common attributes: R1∩R2={P,T}.
Since P→YX⟹P→X, it follows that PT→X. Hence, {P,T} is a superkey of R2.
Thus, R1⋈R2=R12(P,Q,S,T,X) is a lossless join.
Join R12(P,Q,S,T,X) and R34(Q,Y,Z,W):
Common attribute: R12∩R34={Q}.
Attribute closure of Q under the given FDs: Q+={Q,Y,Z,W,X}.
Therefore, Q→{Q,Y,Z,W}, meaning Q is a superkey of R34.
Thus, R12⋈R34=R(P,Q,S,T,X,Y,Z,W) is lossless.
Hence, D1 is a lossless decomposition.
---
### Analysis of Decomposition D2:Constituent subrelations: $R_1(P, Q, S)$, $R_2(T, X)$, $R_3(Q, Y)$, $R_4(Y, Z, W)$.
Notice attribute T appears only in R2(T,X).
The other relations contain attributes {P,Q,S,Y,Z,W}.
The intersection of $R_2(T, X)$ with the union of all other relations is $\{X\}$.
For a lossless join with R2, either X→T or X→(P,Q,S,Y,Z,W) must hold. However, X+={X}, so X cannot determine T or the other attributes.
Hence, D2 is a lossy decomposition.
Therefore, A is the correct option.